In robotics, a robotic paradigm is a mental model of how a robot operates. A robotic paradigm can be described by the relationship between the three primitives of robotics: Sense Plan Act. It can also be described by how sensory data is processed and distributed through the system, and where decisions are made.
Hierarchical/deliberative paradigm
The robot operates in a top-down fashion, heavy on planning.
The robot senses the world, plans the next action, acts; at each step the robot explicitly plans the next move.
All the sensing data tends to be gathered into one global world model.
The reactive paradigm
Sense-act type of organization.
The robot has multiple instances of Sense-Act couplings.
These couplings are concurrent processes, called behaviours, which take the local sensing data and compute the best action to take independently of what the other processes are doing.
The robot will do a combination of behaviours.
Hybrid deliberate/reactive paradigm
The robot first plans (deliberates) how to best decompose a task into subtasks (also called “mission planning”) and then what are the suitable behaviours to accomplish each subtask.
Then the behaviours starts executing as per the Reactive Paradigm.
Sensing organization is also a mixture of Hierarchical and Reactive styles; sensor data gets routed to each behaviour that needs that sensor, but is also available to the planner for construction of a task-oriented global world model.
Source from Wikipedia
2018年11月15日星期四
Reactive planning
In artificial intelligence, reactive planning denotes a group of techniques for action selection by autonomous agents. These techniques differ from classical planning in two aspects. First, they operate in a timely fashion and hence can cope with highly dynamic and unpredictable environments. Second, they compute just one next action in every instant, based on the current context. Reactive planners often (but not always) exploit reactive plans, which are stored structures describing the agent's priorities and behaviour.
Although the term reactive planning goes back to at least 1988, the term "reactive" has now become a pejorative used as an antonym for proactive. Since nearly all agents using reactive planning are proactive, some researchers have begun referring to reactive planning as dynamic planning.
Reactive plan representation
There are several ways to represent a reactive plan. All require a basic representational unit and a means to compose these units into plans.
Condition-action rules (productions)
A condition action rule, or if-then rule, is a rule in the form: if condition then action. These rules are called productions. The meaning of the rule is as follows: if the condition holds, perform the action. The action can be either external (e.g., pick something up and move it), or internal (e.g., write a fact into the internal memory, or evaluate a new set of rules). Conditions are normally boolean and the action either can be performed, or not.
Production rules may be organized in relatively flat structures, but more often are organized into a hierarchy of some kind. For example, subsumption architecture consists of layers of interconnected behaviors, each actually a finite state machine which acts in response to an appropriate input. These layers are then organized into a simple stack, with higher layers subsuming the goals of the lower ones. Other systems may use trees, or may include special mechanisms for changing which goal / rule subset is currently most important. Flat structures are relatively easy to build, but allow only for description of simple behavior, or require immensely complicated conditions to compensate for the lacking structure.
An important part of any distributed action selection algorithms is a conflict resolution mechanism. This is a mechanism for resolving conflicts between actions proposed when more than one rules' condition holds in a given instant. The conflict can be solved for example by
assigning fixed priorities to the rules in advance,
assigning preferences (e.g. in Soar architecture),
learning relative utilities between rules (e.g. in ACT-R),
exploiting a form of planning.
Expert systems often use other simpler heuristics such as recency for selecting rules, but it is difficult to guarantee good behavior in a large system with simple approaches.
Conflict resolution is only necessary for rules that want to take mutually exclusive actions (c.f. Blumberg 1996).
Some limitations of this kind of reactive planning can be found in Brom (2005).
Finite State Machines
Finite state machine (FSM) is model of behaviour of a system. FSMs are used widely in computer science. Modeling behaviour of agents is only one of their possible applications. A typical FSM, when used for describing behaviour of an agent, consists of a set of states and transitions between these states. The transitions are actually condition action rules. In every instant, just one state of the FSM is active, and its transitions are evaluated. If a transition is taken it activates another state. That means, in general transitions are the rules in the following form: if condition then activate-new-state. But transitions can also connect to the 'self' state in some systems, to allow execution of transition actions without actually changing the state.
There are two ways of how to produce behaviour by a FSM. They depend on what is associated with the states by a designer --- they can be either 'acts', or scripts. An 'act' is an atomic action that should be performed by the agent if its FSM is the given state. This action is performed in every time step then. However, more often is the latter case. Here, every state is associated with a script, which describes a sequence of actions that the agent has to perform if its FSM is in a given state. If a transition activates a new state, the former script is simply interrupted, and the new one is started.
If a script is more complicated, it can be broken down to several scripts and a hierarchical FSM can be exploited. In such an automaton, every state can contain substates. Only the states at the atomic level are associated with a script (which is not complicated) or an atomic action.
Computationally, hierarchical FSMs are equivalent to FSMs. That means that each hierarchical FSM can be converted to a classical FSM. However, hierarchical approaches facilitate designs better. See the paper of Damian Isla (2005) for an example of ASM of computer game bots, which uses hierarchical FSMs.
Fuzzy approaches
Both if-then rules and FSMs can be combined with fuzzy logic. The conditions, states and actions are no more boolean or "yes/no" respectively but are approximate and smooth. Consequently, resulted behaviour will transition smoother, especially in the case of transitions between two tasks. However, evaluation of the fuzzy conditions is much slower than evaluation of their crisp counterparts.
See the architecture of Alex Champandard.
Connectionists approaches
Reactive plans can be expressed also by connectionist networks like artificial neural networks or free-flow hierarchies. The basic representational unit is a unit with several input links that feed the unit with "an abstract activity" and output links that propagate the activity to following units. Each unit itself works as the activity transducer. Typically, the units are connected in a layered structure.
Positives of connectionist networks is, first, that the resulted behaviour is more smooth than behaviour produced by crisp if-then rules and FSMs, second, the networks are often adaptive, and third, mechanism of inhibition can be used and hence, behaviour can be also described proscriptively (by means of rules one can describe behaviour only prescriptively). However, the methods have also several flaws. First, for a designer, it is much more complicated to describe behaviour by a network comparing with if-then rules. Second, only relatively simple behaviour can be described, especially if adaptive feature is to be exploited.
Reactive planning algorithms
Typical reactive planning algorithm just evaluates if-then rules or computes the state of a connectionist network. However, some algorithms have special features.
Rete evaluation: with a proper logic representation (which is suitable only for crisp rules), the rules need not to be re-evaluated at every time step. Instead, a form of a cache storing the evaluation from the previous step can be used.
Scripting languages: Sometimes, the rules or FSMs are directly the primitives of an architecture (e.g. in Soar). But more often, reactive plans are programmed in a scripting language, where the rules are only one of the primitives (like in JAM or ABL).
Steering
Steering is a special reactive technique used in navigation of agents. The simplest form of reactive steering is employed in Braitenberg vehicles, which map sensor inputs directly to effector outputs, and can follow or avoid. More complex systems are based on a superposition of attractive or repulsive forces that effect on the agent. This kind of steering is based on the original work on boids of Craig Reynolds. By means of steering, one can achieve a simple form of:
towards a goal navigation
obstacles avoidance behaviour
a wall following behaviour
enemy approaching
predator avoidance
crowd behaviour
The advantage of steering is that it is computationally very efficient. In computer games, hundreds of soldiers can be driven by this technique. In cases of more complicated terrain (e.g. a building), however, steering must be combined with path-finding (as e.g. in Milani ), which is a form of planning .
Source from Wikipedia
Although the term reactive planning goes back to at least 1988, the term "reactive" has now become a pejorative used as an antonym for proactive. Since nearly all agents using reactive planning are proactive, some researchers have begun referring to reactive planning as dynamic planning.
Reactive plan representation
There are several ways to represent a reactive plan. All require a basic representational unit and a means to compose these units into plans.
Condition-action rules (productions)
A condition action rule, or if-then rule, is a rule in the form: if condition then action. These rules are called productions. The meaning of the rule is as follows: if the condition holds, perform the action. The action can be either external (e.g., pick something up and move it), or internal (e.g., write a fact into the internal memory, or evaluate a new set of rules). Conditions are normally boolean and the action either can be performed, or not.
Production rules may be organized in relatively flat structures, but more often are organized into a hierarchy of some kind. For example, subsumption architecture consists of layers of interconnected behaviors, each actually a finite state machine which acts in response to an appropriate input. These layers are then organized into a simple stack, with higher layers subsuming the goals of the lower ones. Other systems may use trees, or may include special mechanisms for changing which goal / rule subset is currently most important. Flat structures are relatively easy to build, but allow only for description of simple behavior, or require immensely complicated conditions to compensate for the lacking structure.
An important part of any distributed action selection algorithms is a conflict resolution mechanism. This is a mechanism for resolving conflicts between actions proposed when more than one rules' condition holds in a given instant. The conflict can be solved for example by
assigning fixed priorities to the rules in advance,
assigning preferences (e.g. in Soar architecture),
learning relative utilities between rules (e.g. in ACT-R),
exploiting a form of planning.
Expert systems often use other simpler heuristics such as recency for selecting rules, but it is difficult to guarantee good behavior in a large system with simple approaches.
Conflict resolution is only necessary for rules that want to take mutually exclusive actions (c.f. Blumberg 1996).
Some limitations of this kind of reactive planning can be found in Brom (2005).
Finite State Machines
Finite state machine (FSM) is model of behaviour of a system. FSMs are used widely in computer science. Modeling behaviour of agents is only one of their possible applications. A typical FSM, when used for describing behaviour of an agent, consists of a set of states and transitions between these states. The transitions are actually condition action rules. In every instant, just one state of the FSM is active, and its transitions are evaluated. If a transition is taken it activates another state. That means, in general transitions are the rules in the following form: if condition then activate-new-state. But transitions can also connect to the 'self' state in some systems, to allow execution of transition actions without actually changing the state.
There are two ways of how to produce behaviour by a FSM. They depend on what is associated with the states by a designer --- they can be either 'acts', or scripts. An 'act' is an atomic action that should be performed by the agent if its FSM is the given state. This action is performed in every time step then. However, more often is the latter case. Here, every state is associated with a script, which describes a sequence of actions that the agent has to perform if its FSM is in a given state. If a transition activates a new state, the former script is simply interrupted, and the new one is started.
If a script is more complicated, it can be broken down to several scripts and a hierarchical FSM can be exploited. In such an automaton, every state can contain substates. Only the states at the atomic level are associated with a script (which is not complicated) or an atomic action.
Computationally, hierarchical FSMs are equivalent to FSMs. That means that each hierarchical FSM can be converted to a classical FSM. However, hierarchical approaches facilitate designs better. See the paper of Damian Isla (2005) for an example of ASM of computer game bots, which uses hierarchical FSMs.
Fuzzy approaches
Both if-then rules and FSMs can be combined with fuzzy logic. The conditions, states and actions are no more boolean or "yes/no" respectively but are approximate and smooth. Consequently, resulted behaviour will transition smoother, especially in the case of transitions between two tasks. However, evaluation of the fuzzy conditions is much slower than evaluation of their crisp counterparts.
See the architecture of Alex Champandard.
Connectionists approaches
Reactive plans can be expressed also by connectionist networks like artificial neural networks or free-flow hierarchies. The basic representational unit is a unit with several input links that feed the unit with "an abstract activity" and output links that propagate the activity to following units. Each unit itself works as the activity transducer. Typically, the units are connected in a layered structure.
Positives of connectionist networks is, first, that the resulted behaviour is more smooth than behaviour produced by crisp if-then rules and FSMs, second, the networks are often adaptive, and third, mechanism of inhibition can be used and hence, behaviour can be also described proscriptively (by means of rules one can describe behaviour only prescriptively). However, the methods have also several flaws. First, for a designer, it is much more complicated to describe behaviour by a network comparing with if-then rules. Second, only relatively simple behaviour can be described, especially if adaptive feature is to be exploited.
Reactive planning algorithms
Typical reactive planning algorithm just evaluates if-then rules or computes the state of a connectionist network. However, some algorithms have special features.
Rete evaluation: with a proper logic representation (which is suitable only for crisp rules), the rules need not to be re-evaluated at every time step. Instead, a form of a cache storing the evaluation from the previous step can be used.
Scripting languages: Sometimes, the rules or FSMs are directly the primitives of an architecture (e.g. in Soar). But more often, reactive plans are programmed in a scripting language, where the rules are only one of the primitives (like in JAM or ABL).
Steering
Steering is a special reactive technique used in navigation of agents. The simplest form of reactive steering is employed in Braitenberg vehicles, which map sensor inputs directly to effector outputs, and can follow or avoid. More complex systems are based on a superposition of attractive or repulsive forces that effect on the agent. This kind of steering is based on the original work on boids of Craig Reynolds. By means of steering, one can achieve a simple form of:
towards a goal navigation
obstacles avoidance behaviour
a wall following behaviour
enemy approaching
predator avoidance
crowd behaviour
The advantage of steering is that it is computationally very efficient. In computer games, hundreds of soldiers can be driven by this technique. In cases of more complicated terrain (e.g. a building), however, steering must be combined with path-finding (as e.g. in Milani ), which is a form of planning .
Source from Wikipedia
2018年11月14日星期三
Developmental robotics
Developmental robotics (DevRob), sometimes called epigenetic robotics, is a scientific field which aims at studying the developmental mechanisms, architectures and constraints that allow lifelong and open-ended learning of new skills and new knowledge in embodied machines. As in human children, learning is expected to be cumulative and of progressively increasing complexity, and to result from self-exploration of the world in combination with social interaction. The typical methodological approach consists in starting from theories of human and animal development elaborated in fields such as developmental psychology, neuroscience, developmental and evolutionary biology, and linguistics, then to formalize and implement them in robots, sometimes exploring extensions or variants of them. The experimentation of those models in robots allows researchers to confront them with reality, and as a consequence developmental robotics also provides feedback and novel hypotheses on theories of human and animal development.
Developmental robotics is related to, but differs from, evolutionary robotics (ER). ER uses populations of robots that evolve over time, whereas DevRob is interested in how the organization of a single robot's control system develops through experience, over time.
DevRob is also related to work done in the domains of robotics and artificial life.
Background
Can a robot learn like a child? Can it learn a variety of new skills and new knowledge unspecified at design time and in a partially unknown and changing environment? How can it discover its body and its relationships with the physical and social environment? How can its cognitive capacities continuously develop without the intervention of an engineer once it is "out of the factory"? What can it learn through natural social interactions with humans? These are the questions at the center of developmental robotics. Alan Turing, as well as a number of other pioneers of cybernetics, already formulated those questions and the general approach in 1950, but it is only since the end of the 20th century that they began to be investigated systematically.
Because the concept of adaptive intelligent machine is central to developmental robotics, it has relationships with fields such as artificial intelligence, machine learning, cognitive robotics or computational neuroscience. Yet, while it may reuse some of the techniques elaborated in these fields, it differs from them from many perspectives. It differs from classical artificial intelligence because it does not assume the capability of advanced symbolic reasoning and focuses on embodied and situated sensorimotor and social skills rather than on abstract symbolic problems. It differs from traditional machine learning because it targets task- independent self-determined learning rather than task-specific inference over "spoon fed human-edited sensori data" (Weng et al., 2001). It differs from cognitive robotics because it focuses on the processes that allow the formation of cognitive capabilities rather than these capabilities themselves. It differs from computational neuroscience because it focuses on functional modeling of integrated architectures of development and learning. More generally, developmental robotics is uniquely characterized by the following three features:
It targets task-independent architectures and learning mechanisms, i.e. the machine/robot has to be able to learn new tasks that are unknown by the engineer;
It emphasizes open-ended development and lifelong learning, i.e. the capacity of an organism to acquire continuously novel skills. This should not be understood as a capacity for learning "anything" or even “everything”, but just that the set of skills that is acquired can be infinitely extended at least in some (not all) directions;
The complexity of acquired knowledge and skills shall increase (and the increase be controlled) progressively.
Developmental robotics emerged at the crossroads of several research communities including embodied artificial intelligence, enactive and dynamical systems cognitive science, connectionism. Starting from the essential idea that learning and development happen as the self-organized result of the dynamical interactions among brains, bodies and their physical and social environment, and trying to understand how this self- organization can be harnessed to provide task-independent lifelong learning of skills of increasing complexity, developmental robotics strongly interacts with fields such as developmental psychology, developmental and cognitive neuroscience, developmental biology (embryology), evolutionary biology, and cognitive linguistics. As many of the theories coming from these sciences are verbal and/or descriptive, this implies a crucial formalization and computational modeling activity in developmental robotics. These computational models are then not only used as ways to explore how to build more versatile and adaptive machines, but also as a way to evaluate their coherence and possibly explore alternative explanations for understanding biological development.
Research directions
Skill domains
Due to the general approach and methodology, developmental robotics projects typically focus on having robots develop the same types of skills as human infants. A first category that is importantly being investigated is the acquisition of sensorimotor skills. These include the discovery of one's own body, including its structure and dynamics such as hand–eye coordination, locomotion, and interaction with objects as well as tool use, with a particular focus on the discovery and learning of affordances. A second category of skills targeted by developmental robots are social and linguistic skills: the acquisition of simple social behavioural games such as turn-taking, coordinated interaction, lexicons, syntax and grammar, and the grounding of these linguistic skills into sensorimotor skills (sometimes referred as symbol grounding). In parallel, the acquisition of associated cognitive skills are being investigated such as the emergence of the self/non-self distinction, the development of attentional capabilities, of categorization systems and higher-level representations of affordances or social constructs, of the emergence of values, empathy, or theories of mind.
Mechanisms and constraints
The sensorimotor and social spaces in which humans and robot live are so large and complex that only a small part of potentially learnable skills can actually be explored and learnt within a life-time. Thus, mechanisms and constraints are necessary to guide developmental organisms in their development and control of the growth of complexity. There are several important families of these guiding mechanisms and constraints which are studied in developmental robotics, all inspired by human development:
Motivational systems, generating internal reward signals that drive exploration and learning, which can be of two main types:
extrinsic motivations push robots/organisms to maintain basic specific internal properties such as food and water level, physical integrity, or light (e.g. in phototropic systems);
intrinsic motivations push robot to search for novelty, challenge, compression or learning progress per se, thus generating what is sometimes called curiosity-driven learning and exploration, or alternatively active learning and exploration;
Social guidance: as humans learn a lot by interacting with their peers, developmental robotics investigates mechanisms which can allow robots to participate to human-like social interaction. By perceiving and interpreting social cues, this may allow robots both to learn from humans (through diverse means such as imitation, emulation, stimulus enhancement, demonstration, etc. ...) and to trigger natural human pedagogy. Thus, social acceptance of developmental robots is also investigated;
Statistical inference biases and cumulative knowledge/skill reuse: biases characterizing both representations/encodings and inference mechanisms can typically allow considerable improvement of the efficiency of learning and are thus studied. Related to this, mechanisms allowing to infer new knowledge and acquire new skills by reusing previously learnt structures is also an essential field of study;
The properties of embodiment, including geometry, materials, or innate motor primitives/synergies often encoded as dynamical systems, can considerably simplify the acquisition of sensorimotor or social skills, and is sometimes referred as morphological computation. The interaction of these constraints with other constraints is an important axis of investigation;
Maturational constraints: In human infants, both the body and the neural system grow progressively, rather than being full-fledged already at birth. This implies for example that new degrees of freedom, as well as increases of the volume and resolution of available sensorimotor signals, may appear as learning and development unfold. Transposing these mechanisms in developmental robots, and understanding how it may hinder or on the contrary ease the acquisition of novel complex skills is a central question in developmental robotics.
From bio-mimetic development to functional inspiration.
While most developmental robotics projects strongly interact with theories of animal and human development, the degrees of similarities and inspiration between identified biological mechanisms and their counterpart in robots, as well as the abstraction levels of modeling, may vary a lot. While some projects aim at modeling precisely both the function and biological implementation (neural or morphological models), such as in neurorobotics, some other projects only focus on functional modeling of the mechanisms and constraints described above, and might for example reuse in their architectures techniques coming from applied mathematics or engineering fields.
Open questions
As developmental robotics is a relatively novel research field and at the same time very ambitious, many fundamental open challenges remain to be solved.
First of all, existing techniques are far from allowing real-world high-dimensional robots to learn an open- ended repertoire of increasingly complex skills over a life-time period. High-dimensional continuous sensorimotor spaces are a major obstacle to be solved. Lifelong cumulative learning is another one. Actually, no experiments lasting more than a few days have been set up so far, which contrasts severely with the time period needed by human infants to learn basic sensorimotor skills while equipped with brains and morphologies which are tremendously more powerful than existing computational mechanisms.
Among the strategies to explore in order to progress towards this target, the interaction between the mechanisms and constraints described in the previous section shall be investigated more systematically. Indeed, they have so far mainly been studied in isolation. For example, the interaction of intrinsically motivated learning and socially guided learning, possibly constrained by maturation, is an essential issue to be investigated.
Another important challenge is to allow robots to perceive, interpret and leverage the diversity of multimodal social cues provided by non-engineer humans during human-robot interaction. These capacities are so far mostly too limited to allow efficient general purpose teaching from humans.
A fundamental scientific issue to be understood and resolved, which applied equally to human development, is how compositionality, functional hierarchies, primitives, and modularity, at all levels of sensorimotor and social structures, can be formed and leveraged during development. This is deeply linked with the problem of the emergence of symbols, sometimes referred as the "symbol grounding problem" when it comes to language acquisition. Actually, the very existence and need for symbols in the brain is actively questioned, and alternative concepts, still allowing for compositionality and functional hierarchies are being investigated.
During biological epigenesis, morphology is not fixed but rather develops in constant interaction with the development of sensorimotor and social skills. The development of morphology poses obvious practical problems with robots, but it may be a crucial mechanism that should be further explored, at least in simulation, such as in morphogenetic robotics.
Another open problem is the understanding of the relation between the key phenomena investigated by developmental robotics (e.g., hierarchical and modular sensorimotor systems, intrinsic/extrinsic/social motivations, and open-ended learning) and the underlying brain mechanisms.
Similarly, in biology, developmental mechanisms (operating at the ontogenetic time scale) strongly interact with evolutionary mechanisms (operating at the phylogenetic time scale) as shown in the flourishing "evo-devo" scientific literature. However, the interaction of those mechanisms in artificial organisms, developmental robots in particular, is still vastly understudied. The interaction of evolutionary mechanisms, unfolding morphologies and developing sensorimotor and social skills will thus be a highly stimulating topic for the future of developmental robotics.
Source from Wikipedia
Developmental robotics is related to, but differs from, evolutionary robotics (ER). ER uses populations of robots that evolve over time, whereas DevRob is interested in how the organization of a single robot's control system develops through experience, over time.
DevRob is also related to work done in the domains of robotics and artificial life.
Background
Can a robot learn like a child? Can it learn a variety of new skills and new knowledge unspecified at design time and in a partially unknown and changing environment? How can it discover its body and its relationships with the physical and social environment? How can its cognitive capacities continuously develop without the intervention of an engineer once it is "out of the factory"? What can it learn through natural social interactions with humans? These are the questions at the center of developmental robotics. Alan Turing, as well as a number of other pioneers of cybernetics, already formulated those questions and the general approach in 1950, but it is only since the end of the 20th century that they began to be investigated systematically.
Because the concept of adaptive intelligent machine is central to developmental robotics, it has relationships with fields such as artificial intelligence, machine learning, cognitive robotics or computational neuroscience. Yet, while it may reuse some of the techniques elaborated in these fields, it differs from them from many perspectives. It differs from classical artificial intelligence because it does not assume the capability of advanced symbolic reasoning and focuses on embodied and situated sensorimotor and social skills rather than on abstract symbolic problems. It differs from traditional machine learning because it targets task- independent self-determined learning rather than task-specific inference over "spoon fed human-edited sensori data" (Weng et al., 2001). It differs from cognitive robotics because it focuses on the processes that allow the formation of cognitive capabilities rather than these capabilities themselves. It differs from computational neuroscience because it focuses on functional modeling of integrated architectures of development and learning. More generally, developmental robotics is uniquely characterized by the following three features:
It targets task-independent architectures and learning mechanisms, i.e. the machine/robot has to be able to learn new tasks that are unknown by the engineer;
It emphasizes open-ended development and lifelong learning, i.e. the capacity of an organism to acquire continuously novel skills. This should not be understood as a capacity for learning "anything" or even “everything”, but just that the set of skills that is acquired can be infinitely extended at least in some (not all) directions;
The complexity of acquired knowledge and skills shall increase (and the increase be controlled) progressively.
Developmental robotics emerged at the crossroads of several research communities including embodied artificial intelligence, enactive and dynamical systems cognitive science, connectionism. Starting from the essential idea that learning and development happen as the self-organized result of the dynamical interactions among brains, bodies and their physical and social environment, and trying to understand how this self- organization can be harnessed to provide task-independent lifelong learning of skills of increasing complexity, developmental robotics strongly interacts with fields such as developmental psychology, developmental and cognitive neuroscience, developmental biology (embryology), evolutionary biology, and cognitive linguistics. As many of the theories coming from these sciences are verbal and/or descriptive, this implies a crucial formalization and computational modeling activity in developmental robotics. These computational models are then not only used as ways to explore how to build more versatile and adaptive machines, but also as a way to evaluate their coherence and possibly explore alternative explanations for understanding biological development.
Research directions
Skill domains
Due to the general approach and methodology, developmental robotics projects typically focus on having robots develop the same types of skills as human infants. A first category that is importantly being investigated is the acquisition of sensorimotor skills. These include the discovery of one's own body, including its structure and dynamics such as hand–eye coordination, locomotion, and interaction with objects as well as tool use, with a particular focus on the discovery and learning of affordances. A second category of skills targeted by developmental robots are social and linguistic skills: the acquisition of simple social behavioural games such as turn-taking, coordinated interaction, lexicons, syntax and grammar, and the grounding of these linguistic skills into sensorimotor skills (sometimes referred as symbol grounding). In parallel, the acquisition of associated cognitive skills are being investigated such as the emergence of the self/non-self distinction, the development of attentional capabilities, of categorization systems and higher-level representations of affordances or social constructs, of the emergence of values, empathy, or theories of mind.
Mechanisms and constraints
The sensorimotor and social spaces in which humans and robot live are so large and complex that only a small part of potentially learnable skills can actually be explored and learnt within a life-time. Thus, mechanisms and constraints are necessary to guide developmental organisms in their development and control of the growth of complexity. There are several important families of these guiding mechanisms and constraints which are studied in developmental robotics, all inspired by human development:
Motivational systems, generating internal reward signals that drive exploration and learning, which can be of two main types:
extrinsic motivations push robots/organisms to maintain basic specific internal properties such as food and water level, physical integrity, or light (e.g. in phototropic systems);
intrinsic motivations push robot to search for novelty, challenge, compression or learning progress per se, thus generating what is sometimes called curiosity-driven learning and exploration, or alternatively active learning and exploration;
Social guidance: as humans learn a lot by interacting with their peers, developmental robotics investigates mechanisms which can allow robots to participate to human-like social interaction. By perceiving and interpreting social cues, this may allow robots both to learn from humans (through diverse means such as imitation, emulation, stimulus enhancement, demonstration, etc. ...) and to trigger natural human pedagogy. Thus, social acceptance of developmental robots is also investigated;
Statistical inference biases and cumulative knowledge/skill reuse: biases characterizing both representations/encodings and inference mechanisms can typically allow considerable improvement of the efficiency of learning and are thus studied. Related to this, mechanisms allowing to infer new knowledge and acquire new skills by reusing previously learnt structures is also an essential field of study;
The properties of embodiment, including geometry, materials, or innate motor primitives/synergies often encoded as dynamical systems, can considerably simplify the acquisition of sensorimotor or social skills, and is sometimes referred as morphological computation. The interaction of these constraints with other constraints is an important axis of investigation;
Maturational constraints: In human infants, both the body and the neural system grow progressively, rather than being full-fledged already at birth. This implies for example that new degrees of freedom, as well as increases of the volume and resolution of available sensorimotor signals, may appear as learning and development unfold. Transposing these mechanisms in developmental robots, and understanding how it may hinder or on the contrary ease the acquisition of novel complex skills is a central question in developmental robotics.
From bio-mimetic development to functional inspiration.
While most developmental robotics projects strongly interact with theories of animal and human development, the degrees of similarities and inspiration between identified biological mechanisms and their counterpart in robots, as well as the abstraction levels of modeling, may vary a lot. While some projects aim at modeling precisely both the function and biological implementation (neural or morphological models), such as in neurorobotics, some other projects only focus on functional modeling of the mechanisms and constraints described above, and might for example reuse in their architectures techniques coming from applied mathematics or engineering fields.
Open questions
As developmental robotics is a relatively novel research field and at the same time very ambitious, many fundamental open challenges remain to be solved.
First of all, existing techniques are far from allowing real-world high-dimensional robots to learn an open- ended repertoire of increasingly complex skills over a life-time period. High-dimensional continuous sensorimotor spaces are a major obstacle to be solved. Lifelong cumulative learning is another one. Actually, no experiments lasting more than a few days have been set up so far, which contrasts severely with the time period needed by human infants to learn basic sensorimotor skills while equipped with brains and morphologies which are tremendously more powerful than existing computational mechanisms.
Among the strategies to explore in order to progress towards this target, the interaction between the mechanisms and constraints described in the previous section shall be investigated more systematically. Indeed, they have so far mainly been studied in isolation. For example, the interaction of intrinsically motivated learning and socially guided learning, possibly constrained by maturation, is an essential issue to be investigated.
Another important challenge is to allow robots to perceive, interpret and leverage the diversity of multimodal social cues provided by non-engineer humans during human-robot interaction. These capacities are so far mostly too limited to allow efficient general purpose teaching from humans.
A fundamental scientific issue to be understood and resolved, which applied equally to human development, is how compositionality, functional hierarchies, primitives, and modularity, at all levels of sensorimotor and social structures, can be formed and leveraged during development. This is deeply linked with the problem of the emergence of symbols, sometimes referred as the "symbol grounding problem" when it comes to language acquisition. Actually, the very existence and need for symbols in the brain is actively questioned, and alternative concepts, still allowing for compositionality and functional hierarchies are being investigated.
During biological epigenesis, morphology is not fixed but rather develops in constant interaction with the development of sensorimotor and social skills. The development of morphology poses obvious practical problems with robots, but it may be a crucial mechanism that should be further explored, at least in simulation, such as in morphogenetic robotics.
Another open problem is the understanding of the relation between the key phenomena investigated by developmental robotics (e.g., hierarchical and modular sensorimotor systems, intrinsic/extrinsic/social motivations, and open-ended learning) and the underlying brain mechanisms.
Similarly, in biology, developmental mechanisms (operating at the ontogenetic time scale) strongly interact with evolutionary mechanisms (operating at the phylogenetic time scale) as shown in the flourishing "evo-devo" scientific literature. However, the interaction of those mechanisms in artificial organisms, developmental robots in particular, is still vastly understudied. The interaction of evolutionary mechanisms, unfolding morphologies and developing sensorimotor and social skills will thus be a highly stimulating topic for the future of developmental robotics.
Source from Wikipedia
Adaptable robotics
Adaptable robotics are generally based in robot developer kits. This technology is distinguished from static automation due to its capacity to adapt to changing environmental conditions and material features while retaining a degree of predictability required for collaboration (e.g. human-robot collaboration). The degree of adaptability is demonstrated in the way these can be moved around and used in different tasks. Unlike static or factory robots, which have pre-defined way of operating, adaptable robots can function even if a component breaks, making them useful in cases like caring for the elderly, doing household tasks, and rescue work.
The kits come with an open software platform tailored to a range of common robotic functions. The kits also come with common robotics hardware that connects easily with the software (infrared sensors, motors, microphone and video camera), which add to the capabilities of the robot. The process of modifying a robot to achieve varying capabilities such as collaboration could merely include the selection of a module, the exchange of modules, robotic instruction via software, and execution.
Robot kit
A robot kit is a special construction kit for building robots, especially autonomous mobile robots.
Toy robot kits are also supplied by several companies. They are mostly made of plastics elements like Lego Mindstorms, rero Reconfigurable Robot kit, the Robotis Bioloid, Robobuilder, the ROBO-BOX-3.0 (produced by Inex), and the lesser-known KAI Robot (produced by Kaimax), or aluminium elements like Lynxmotion's Servo Erector Set and the qfix kit.
The kits can consist of: structural elements, mechanical elements, motors (or other actuators), sensors and a controller board to control the inputs and outputs of the robot. In some cases, the kits can be available without electronics as well, to provide the user the opportunity to use his or her own.
Robot kits
Arduino controlling Tamiya (or another) kit
Lego Mindstorms
Lynxmotion
Qfix robot kit
rero Reconfigurable Robot kit
Robotis Bioloid
Stiquito
Tetrix Robotics Kit
Vex Robotics Design System
Vorpal The Hexapod
WonderBorg
Vex
MyRobotTime
Source from Wikipedia
The kits come with an open software platform tailored to a range of common robotic functions. The kits also come with common robotics hardware that connects easily with the software (infrared sensors, motors, microphone and video camera), which add to the capabilities of the robot. The process of modifying a robot to achieve varying capabilities such as collaboration could merely include the selection of a module, the exchange of modules, robotic instruction via software, and execution.
Robot kit
A robot kit is a special construction kit for building robots, especially autonomous mobile robots.
Toy robot kits are also supplied by several companies. They are mostly made of plastics elements like Lego Mindstorms, rero Reconfigurable Robot kit, the Robotis Bioloid, Robobuilder, the ROBO-BOX-3.0 (produced by Inex), and the lesser-known KAI Robot (produced by Kaimax), or aluminium elements like Lynxmotion's Servo Erector Set and the qfix kit.
The kits can consist of: structural elements, mechanical elements, motors (or other actuators), sensors and a controller board to control the inputs and outputs of the robot. In some cases, the kits can be available without electronics as well, to provide the user the opportunity to use his or her own.
Robot kits
Arduino controlling Tamiya (or another) kit
Lego Mindstorms
Lynxmotion
Qfix robot kit
rero Reconfigurable Robot kit
Robotis Bioloid
Stiquito
Tetrix Robotics Kit
Vex Robotics Design System
Vorpal The Hexapod
WonderBorg
Vex
MyRobotTime
Source from Wikipedia
Robot software
Robot software is the set of coded commands or instructions that tell a mechanical device and electronic system, known together as a robot, what tasks to perform. Robot software is used to perform autonomous tasks. Many software systems and frameworks have been proposed to make programming robots easier.
Some robot software aims at developing intelligent mechanical devices. Common tasks include feedback loops, control, pathfinding, data filtering, locating and sharing data.
Introduction
While it is a specific type of software, it is still quite diverse. Each manufacturer has their own robot software. While the vast majority of software is about manipulation of data and seeing the result on-screen, robot software is for the manipulation of objects or tools in the real world.
Industrial robot software
Software for industrial robots consists of data objects and lists of instructions, known as program flow (list of instructions). For example,
Go to Jig1
is an instruction to the robot to go to positional data named Jig1. Of course programs can also contain implicit data for example
Tell axis 1 move 30 degrees.
Data and program usually reside in separate sections of the robot controller memory. One can change the data without changing the program and vice versa. For example, one can write a different program using the same Jig1 or one can adjust the position of Jig1 without changing the programs that use it.
Examples of programming languages for industrial robots
Due to the highly proprietary nature of robot software, most manufacturers of robot hardware also provide their own software. While this is not unusual in other automated control systems, the lack of standardization of programming methods for robots does pose certain challenges. For example, there are over 30 different manufacturers of industrial robots, so there are also 30 different robot programming languages required. There are enough similarities between the different robots that it is possible to gain a broad-based understanding of robot programming without having to learn each manufacturer's proprietary language.
One method of controlling robots from multiple manufacturers is to use a Post processor and Off-line programming (robotics) software. With this method, it is possible to handle brand-specific robot programming language from a universal programming language, such as Python (programming language). however, compiling and uploading fixed off-line code to a robot controller doesn't allow the robotic system to be state aware, so it cannot adapt it's motion and recover as the environment changes. Unified real-time adaptive control for any robot is currently only possible with Actin.
Some examples of published robot programming languages are shown below.
Task in plain English:
PROGRAM PICKPLACE
Epson RC+ (example for a vacuum pickup)
Clearly the robot should not continue the next move until the gripper is completely closed. Confirmation or allowed time is implicit in the above examples of CLOSEI and GRIP whereas the On vacuum command requires a time delay to ensure satisfactory suction.
Other robot programming languages
Visual programming language
The LEGO Mindstorms EV3 programming language is a simple language for its users to interact with. It is a graphical user interface (GUI) written with LabVIEW. The approach is to start with the program rather than the data. The program is constructed by dragging icons into the program area and adding or inserting into the sequence. For each icon you then specify the parameters (data). For example, for the motor drive icon you specify which motors and by how much they move. When the program is written it is downloaded into the Lego NXT 'brick' (microcontroller) for test.
Scripting languages
A scripting language is a high-level programming language that is used to control the software application, and is interpreted in real-time, or "translated on the fly", instead of being compiled in advance. A scripting language may be a general-purpose programming language or it may be limited to specific functions used to augment the running of an application or system program. Some scripting languages, such as RoboLogix, have data objects residing in registers, and the program flow represents the list of instructions, or instruction set, that is used to program the robot.
Programming languages in industrial robotics
Robot brand Language name
ABB RAPID
Comau PDL2
Fanuc Karel
Kawasaki AS
Kuka KRL
Stäubli VAL3
Yaskawa Inform
Programming languages are generally designed for building data structures and algorithms from scratch, while scripting languages are intended more for connecting, or “gluing”, components and instructions together. Consequently, the scripting language instruction set is usually a streamlined list of program commands that are used to simplify the programming process and provide rapid application development.
Parallel languages
Another interesting approach is worthy of mention. All robotic applications need parallelism and event-based programming. Parallelism is where the robot does two or more things at the same time. This requires appropriate hardware and software. Most programming languages rely on threads or complex abstraction classes to handle parallelism and the complexity that comes with it, like concurrent access to shared resources. URBI provides a higher level of abstraction by integrating parallelism and events in the core of the language semantics.
whenever(face.visible)
{
headPan.val += camera.xfov * face.x
&
headTilt.val += camera.yfov * face.y
}
The above code will move the headPan and headTilt motors in parallel to make the robot head follow the human face visible on the video taken by its camera whenever a face is seen by the robot.
Robot application software
Regardless which language is used, the end result of robot software is to create robotic applications that help or entertain people. Applications include command-and-control and tasking software. Command-and-control software includes robot control GUIs for tele-operated robots, point-n-click command software for autonomous robots, and scheduling software for mobile robots in factories. Tasking software includes simple drag-n-drop interfaces for setting up delivery routes, security patrols and visitor tours; it also includes custom programs written to deploy specific applications. General purpose robot application software is deployed on widely distributed robotic platforms.
Safety considerations
Programming errors represent a serious safety consideration, particularly in large industrial robots. The power and size of industrial robots mean they are capable of inflicting severe injury if programmed incorrectly or used in an unsafe manner. Due to the mass and high-speeds of industrial robots, it is always unsafe for a human to remain in the work area of the robot during automatic operation. The system can begin motion at unexpected times and a human will be unable to react quickly enough in many situations, even if prepared to do so. Thus, even if the software is free of programming errors, great care must be taken to make an industrial robot safe for human workers or human interaction, such as loading or unloading parts, clearing a part jam, or performing maintenance. The ANSI/RIA R15.06-1999 American National Standard for Industrial Robots and Robot Systems - Safety Requirements (revision of ANSI/RIA R15.06-1992) book from the Robotic Industries Association is the accepted standard on robot safety. This includes guidelines for both the design of industrial robots, and the implementation or integration and use of industrial robots on the factory floor. Numerous safety concepts such as safety controllers, maximum speed during a teach mode, and use of physical barriers are covered.
Robotics software projects
The following is a list of open source and free software for robotics projects.
CLARAty - Coupled-Layer Architecture for Robotic Autonomy. It is a collaborative effort among four institutions: NASA Jet Propulsion Laboratory, NASA Ames Research Center, Carnegie Mellon, and the University of Minnesota.
dLife - Free/Open Source Java library for Robotics, AI and Vision. Supports Pioneer, Khepera II & II, Hemission, Aibo and Finch robots as well as Player/Stage simulations.
Experimental Robotics Framework - A software for making experiments with multiple robots in 3d, with support for the latest technologies, that sits on top of Player/Stage and Open/CV.
MARIE - Mobile and Autonomous Robotics Integration Environment - is a Free Software using a component-based approach to build robotics software systems by integrating previously existing and new software components.
Microsoft - Microsoft Robotics Developer Studio
OpenRDK - An open-source software framework for robotics for developing loosely coupled modules. It provides transparent concurrency management, inter-process (via sockets) and intra-process (via shared memory) blackboard-based communication and a linking technique that allows for input/output data ports conceptual system design. Modules for connecting to simulators and generic robot drivers are provided.
OpenRTM-aist - a software platform developed on the basis of the RT middleware standard. It is developed by National Institute of Advanced Industrial Science and Technology in Japan.
OROCOS - the Open Robot Control Software project provides a Free Software toolkit for realtime robot arm and machine tool control.
OPRoS - Open Platform Robotic Services is an open source project for robot development. It provides a solution including robot platform and GUI developing tools with the source code of robot device components.
RoboDK - A robot development kit platform to simulate industrial robots. RoboDK allows you to program any robot using Python and handles brand-specific syntax depending on your robot controller.
Robotics Library is an open-source C++ library covering kinematics, planning, visualization, and hardware drivers for several robots.
Robotics Toolbox for MATLAB - this is Free Software that provides functionality for representing pose (homogeneous transformations, Euler and RPY angles, quaternions), arm robots (forward/inverse kinematics and dynamics) and mobile robots (control, localisation and planning)
Rossum Project, G.W. Lucas,open-source robotics project .
Player/Stage Project - A very popular Free Software robot interface and simulation system, used for robotics research and teaching worldwide.
Pyro, Python Robotics - Popular robotics Free Software used in universities and colleges.
RoboMind - Educational software to learn the basics of robotics and programming.
Robot Operating System - Robot Operating System is an open-source platform for robot programming using Python and C++. Java, Lisp, Lua and Pharo are supported but still in experimental stage.
Robot Intelligence Kernel
Off-line Programming
Visual Components is a commercial 3D discrete event simulation software that enables material flow and robotics simulation on one platform. The functionality can be extended with off-line programming capabilities and PLC connectivity.
"The Basics - Robot Software". Seattle Robotics Society.
"Mobile Autonomous Robot Software (MARS)". Georgia Tech Research Corporation.
"Tech Database". robot.spawar.navy.mil.
Adaptive Robotics Software at the Idaho National Laboratory
On-line Control
Energid Technologies develops and sells the Actin SDK which is a robotics toolkit, optimization engine, and constraint-management system written in C++ that uses Jacobian-based kinematics to optimize performance over all types of robots. Since Actin is sold as an SDK, it can be white labeled or OEMed into other robot control software packages. This approach allows users to develop control of a known system in an offline mode through a GUI and then let Actin handle the high-level online control to stream constantly updated joint-level commands to the robot/controller in real time to avoid collisions by adapting to sensed environmental changes.
Source from Wikipedia
Some robot software aims at developing intelligent mechanical devices. Common tasks include feedback loops, control, pathfinding, data filtering, locating and sharing data.
Introduction
While it is a specific type of software, it is still quite diverse. Each manufacturer has their own robot software. While the vast majority of software is about manipulation of data and seeing the result on-screen, robot software is for the manipulation of objects or tools in the real world.
Industrial robot software
Software for industrial robots consists of data objects and lists of instructions, known as program flow (list of instructions). For example,
Go to Jig1
is an instruction to the robot to go to positional data named Jig1. Of course programs can also contain implicit data for example
Tell axis 1 move 30 degrees.
Data and program usually reside in separate sections of the robot controller memory. One can change the data without changing the program and vice versa. For example, one can write a different program using the same Jig1 or one can adjust the position of Jig1 without changing the programs that use it.
Examples of programming languages for industrial robots
Due to the highly proprietary nature of robot software, most manufacturers of robot hardware also provide their own software. While this is not unusual in other automated control systems, the lack of standardization of programming methods for robots does pose certain challenges. For example, there are over 30 different manufacturers of industrial robots, so there are also 30 different robot programming languages required. There are enough similarities between the different robots that it is possible to gain a broad-based understanding of robot programming without having to learn each manufacturer's proprietary language.
One method of controlling robots from multiple manufacturers is to use a Post processor and Off-line programming (robotics) software. With this method, it is possible to handle brand-specific robot programming language from a universal programming language, such as Python (programming language). however, compiling and uploading fixed off-line code to a robot controller doesn't allow the robotic system to be state aware, so it cannot adapt it's motion and recover as the environment changes. Unified real-time adaptive control for any robot is currently only possible with Actin.
Some examples of published robot programming languages are shown below.
Task in plain English:
Move to P1 (a general safe position)Example program:
Move to P2 (an approach to P3)
Move to P3 (a position to pick the object)
Close gripper
Move to P4 (an approach to P5)
Move to P5 (a position to place the object)
Open gripper
Move to P1 and finish
VAL was one of the first robot ‘languages’ and was used in Unimate robots. Variants of VAL have been used by other manufacturers including Adept Technology. Stäubli currently use VAL3.
PROGRAM PICKPLACE
1. MOVE P1Example of Stäubli VAL3 program:
2. MOVE P2
3. MOVE P3
4. CLOSEI 0.00
5. MOVE P4
6. MOVE P5
7. OPENI 0.00
8. MOVE P1
.END
begintrAppro is cartesian transformation variable. If we use in with appro command, we do not need to teach P2 and P4 point, but we dynamically transform an approach to position of pick and place for trajectory generation.
movej(p1,tGripper,mNomSpeed)
movej(appro(p3,trAppro),tGripper,mNomSpeed)
movel(p3,tGripper,mNomSpeed)
close(tGripper)
movej(appro(p5,trAppro),tGripper,mNomSpeed)
movel(p5,tGripper,mNomSpeed)
open(tGripper)
movej(p1,tGripper,mNomSpeed)
end
Epson RC+ (example for a vacuum pickup)
Function PickPlaceROBOFORTH (a language based on FORTH).
Jump P1
Jump P2
Jump P3
On vacuum
Wait .1
Jump P4
Jump P5
Off vacuum
Wait .1
Jump P1
Fend
: PICKPLACE(With Roboforth you can specify approach positions for places so you do not need P2 and P4.)
P1
P3 GRIP WITHDRAW
P5 UNGRIP WITHDRAW
P1
;
Clearly the robot should not continue the next move until the gripper is completely closed. Confirmation or allowed time is implicit in the above examples of CLOSEI and GRIP whereas the On vacuum command requires a time delay to ensure satisfactory suction.
Other robot programming languages
Visual programming language
The LEGO Mindstorms EV3 programming language is a simple language for its users to interact with. It is a graphical user interface (GUI) written with LabVIEW. The approach is to start with the program rather than the data. The program is constructed by dragging icons into the program area and adding or inserting into the sequence. For each icon you then specify the parameters (data). For example, for the motor drive icon you specify which motors and by how much they move. When the program is written it is downloaded into the Lego NXT 'brick' (microcontroller) for test.
Scripting languages
A scripting language is a high-level programming language that is used to control the software application, and is interpreted in real-time, or "translated on the fly", instead of being compiled in advance. A scripting language may be a general-purpose programming language or it may be limited to specific functions used to augment the running of an application or system program. Some scripting languages, such as RoboLogix, have data objects residing in registers, and the program flow represents the list of instructions, or instruction set, that is used to program the robot.
Programming languages in industrial robotics
Robot brand Language name
ABB RAPID
Comau PDL2
Fanuc Karel
Kawasaki AS
Kuka KRL
Stäubli VAL3
Yaskawa Inform
Programming languages are generally designed for building data structures and algorithms from scratch, while scripting languages are intended more for connecting, or “gluing”, components and instructions together. Consequently, the scripting language instruction set is usually a streamlined list of program commands that are used to simplify the programming process and provide rapid application development.
Parallel languages
Another interesting approach is worthy of mention. All robotic applications need parallelism and event-based programming. Parallelism is where the robot does two or more things at the same time. This requires appropriate hardware and software. Most programming languages rely on threads or complex abstraction classes to handle parallelism and the complexity that comes with it, like concurrent access to shared resources. URBI provides a higher level of abstraction by integrating parallelism and events in the core of the language semantics.
whenever(face.visible)
{
headPan.val += camera.xfov * face.x
&
headTilt.val += camera.yfov * face.y
}
The above code will move the headPan and headTilt motors in parallel to make the robot head follow the human face visible on the video taken by its camera whenever a face is seen by the robot.
Robot application software
Regardless which language is used, the end result of robot software is to create robotic applications that help or entertain people. Applications include command-and-control and tasking software. Command-and-control software includes robot control GUIs for tele-operated robots, point-n-click command software for autonomous robots, and scheduling software for mobile robots in factories. Tasking software includes simple drag-n-drop interfaces for setting up delivery routes, security patrols and visitor tours; it also includes custom programs written to deploy specific applications. General purpose robot application software is deployed on widely distributed robotic platforms.
Safety considerations
Programming errors represent a serious safety consideration, particularly in large industrial robots. The power and size of industrial robots mean they are capable of inflicting severe injury if programmed incorrectly or used in an unsafe manner. Due to the mass and high-speeds of industrial robots, it is always unsafe for a human to remain in the work area of the robot during automatic operation. The system can begin motion at unexpected times and a human will be unable to react quickly enough in many situations, even if prepared to do so. Thus, even if the software is free of programming errors, great care must be taken to make an industrial robot safe for human workers or human interaction, such as loading or unloading parts, clearing a part jam, or performing maintenance. The ANSI/RIA R15.06-1999 American National Standard for Industrial Robots and Robot Systems - Safety Requirements (revision of ANSI/RIA R15.06-1992) book from the Robotic Industries Association is the accepted standard on robot safety. This includes guidelines for both the design of industrial robots, and the implementation or integration and use of industrial robots on the factory floor. Numerous safety concepts such as safety controllers, maximum speed during a teach mode, and use of physical barriers are covered.
Robotics software projects
The following is a list of open source and free software for robotics projects.
CLARAty - Coupled-Layer Architecture for Robotic Autonomy. It is a collaborative effort among four institutions: NASA Jet Propulsion Laboratory, NASA Ames Research Center, Carnegie Mellon, and the University of Minnesota.
dLife - Free/Open Source Java library for Robotics, AI and Vision. Supports Pioneer, Khepera II & II, Hemission, Aibo and Finch robots as well as Player/Stage simulations.
Experimental Robotics Framework - A software for making experiments with multiple robots in 3d, with support for the latest technologies, that sits on top of Player/Stage and Open/CV.
MARIE - Mobile and Autonomous Robotics Integration Environment - is a Free Software using a component-based approach to build robotics software systems by integrating previously existing and new software components.
Microsoft - Microsoft Robotics Developer Studio
OpenRDK - An open-source software framework for robotics for developing loosely coupled modules. It provides transparent concurrency management, inter-process (via sockets) and intra-process (via shared memory) blackboard-based communication and a linking technique that allows for input/output data ports conceptual system design. Modules for connecting to simulators and generic robot drivers are provided.
OpenRTM-aist - a software platform developed on the basis of the RT middleware standard. It is developed by National Institute of Advanced Industrial Science and Technology in Japan.
OROCOS - the Open Robot Control Software project provides a Free Software toolkit for realtime robot arm and machine tool control.
OPRoS - Open Platform Robotic Services is an open source project for robot development. It provides a solution including robot platform and GUI developing tools with the source code of robot device components.
RoboDK - A robot development kit platform to simulate industrial robots. RoboDK allows you to program any robot using Python and handles brand-specific syntax depending on your robot controller.
Robotics Library is an open-source C++ library covering kinematics, planning, visualization, and hardware drivers for several robots.
Robotics Toolbox for MATLAB - this is Free Software that provides functionality for representing pose (homogeneous transformations, Euler and RPY angles, quaternions), arm robots (forward/inverse kinematics and dynamics) and mobile robots (control, localisation and planning)
Rossum Project, G.W. Lucas,open-source robotics project .
Player/Stage Project - A very popular Free Software robot interface and simulation system, used for robotics research and teaching worldwide.
Pyro, Python Robotics - Popular robotics Free Software used in universities and colleges.
RoboMind - Educational software to learn the basics of robotics and programming.
Robot Operating System - Robot Operating System is an open-source platform for robot programming using Python and C++. Java, Lisp, Lua and Pharo are supported but still in experimental stage.
Robot Intelligence Kernel
Off-line Programming
Visual Components is a commercial 3D discrete event simulation software that enables material flow and robotics simulation on one platform. The functionality can be extended with off-line programming capabilities and PLC connectivity.
"The Basics - Robot Software". Seattle Robotics Society.
"Mobile Autonomous Robot Software (MARS)". Georgia Tech Research Corporation.
"Tech Database". robot.spawar.navy.mil.
Adaptive Robotics Software at the Idaho National Laboratory
On-line Control
Energid Technologies develops and sells the Actin SDK which is a robotics toolkit, optimization engine, and constraint-management system written in C++ that uses Jacobian-based kinematics to optimize performance over all types of robots. Since Actin is sold as an SDK, it can be white labeled or OEMed into other robot control software packages. This approach allows users to develop control of a known system in an offline mode through a GUI and then let Actin handle the high-level online control to stream constantly updated joint-level commands to the robot/controller in real time to avoid collisions by adapting to sensed environmental changes.
Source from Wikipedia
Hexapod robotics
A hexapod robot is a mechanical vehicle that walks on six legs. Since a robot can be statically stable on three or more legs, a hexapod robot has a great deal of flexibility in how it can move. If legs become disabled, the robot may still be able to walk. Furthermore, not all of the robot's legs are needed for stability; other legs are free to reach new foot placements or manipulate a payload.
Many hexapod robots are biologically inspired by Hexapoda locomotion. Hexapods may be used to test biological theories about insect locomotion, motor control, and neurobiology.
Designs
Hexapod designs vary in leg arrangement. Insect-inspired robots are typically laterally symmetric, such as the RiSE robot at Carnegie Mellon. A radially symmetric hexapod is ATHLETE (All-Terrain Hex-Legged Extra-Terrestrial Explorer) robot at JPL.
Typically, individual legs range from two to six degrees of freedom. Hexapod feet are typically pointed, but can also be tipped with adhesive material to help climb walls or wheels so the robot can drive quickly when the ground is flat.
Locomotion
Most often, hexapods are controlled by gaits, which allow the robot to move forward, turn, and perhaps side-step. Some of the most common gaits are as follows:
Alternating tripod: 3 legs on the ground at a time.
Quadruped.
Crawl: move just one leg at a time.
Gaits for hexapods are often stable, even in slightly rocky and uneven terrain.
Motion may also be nongaited, which means the sequence of leg motions is not fixed, but rather chosen by the computer in response to the sensed environment. This may be most helpful in very rocky terrain, but existing techniques for motion planning are computationally expensive.
Biologically inspired
Insects are chosen as models because their nervous system are simpler than other animal species. Also, complex behaviours can be attributed to just a few neurons and the pathway between sensory input and motor output is relatively shorter. Insects' walking behaviour and neural architecture are used to improve robot locomotion. Conversely, biologists can use hexapod robots for testing different hypotheses.
Biologically inspired hexapod robots largely depend on the insect species used as a model. The cockroach and the stick insect are the two most commonly used insect species; both have been ethologically and neurophysiologically extensively studied. At present no complete nervous system is known, therefore, models usually combine different insect models, including those of other insects.
Insect gaits are usually obtained by two approaches: the centralized and the decentralized control architectures. Centralized controllers directly specify transitions of all legs, whereas in decentralized architectures, six nodes (legs) are connected in a parallel network; gaits arise by the interaction between neighbouring legs.
Foot coordination
The term "foot coordination" refers to the mechanism responsible for controlling the transition between footsteps; considering that the body does not turn around. Most approaches try to replicate the appearance of known insects, for example the tripod or tetrapod shape. However, other approaches have been used to find stable paces; for example, by launching programs using genetic algorithms or by optimizing the energy of walking.
The walking patterns of insects are usually obtained by two approaches: centralized and decentralized control architectures. The centralized controllers directly specify the transitions of all the legs, whereas in the decentralized architectures, six nodes (legs) are connected in a parallel network; the paces are obtained thanks to the interaction between the neighboring legs.
Controller of the paw
There are no borders to the complexity of the morphology of the paw. However, legs that are built on an insect model usually have between two and six degrees of freedom. Leg segments typically bear the name of their biological counterpart, which are similar for most species. From the body to the end of the paw, the segments bear the names of coxa, femur and tibia ; typically the joints between the coxa and the femur and between the femur and the tibia are considered as simple hinges. The models of the articulation between the body and the coxa comprise between one and three degrees of freedom, depending on the species and the thoracic segment on which the leg is enclosed.
Source from Wikipedia
Many hexapod robots are biologically inspired by Hexapoda locomotion. Hexapods may be used to test biological theories about insect locomotion, motor control, and neurobiology.
Designs
Hexapod designs vary in leg arrangement. Insect-inspired robots are typically laterally symmetric, such as the RiSE robot at Carnegie Mellon. A radially symmetric hexapod is ATHLETE (All-Terrain Hex-Legged Extra-Terrestrial Explorer) robot at JPL.
Typically, individual legs range from two to six degrees of freedom. Hexapod feet are typically pointed, but can also be tipped with adhesive material to help climb walls or wheels so the robot can drive quickly when the ground is flat.
Locomotion
Most often, hexapods are controlled by gaits, which allow the robot to move forward, turn, and perhaps side-step. Some of the most common gaits are as follows:
Alternating tripod: 3 legs on the ground at a time.
Quadruped.
Crawl: move just one leg at a time.
Gaits for hexapods are often stable, even in slightly rocky and uneven terrain.
Motion may also be nongaited, which means the sequence of leg motions is not fixed, but rather chosen by the computer in response to the sensed environment. This may be most helpful in very rocky terrain, but existing techniques for motion planning are computationally expensive.
Biologically inspired
Insects are chosen as models because their nervous system are simpler than other animal species. Also, complex behaviours can be attributed to just a few neurons and the pathway between sensory input and motor output is relatively shorter. Insects' walking behaviour and neural architecture are used to improve robot locomotion. Conversely, biologists can use hexapod robots for testing different hypotheses.
Biologically inspired hexapod robots largely depend on the insect species used as a model. The cockroach and the stick insect are the two most commonly used insect species; both have been ethologically and neurophysiologically extensively studied. At present no complete nervous system is known, therefore, models usually combine different insect models, including those of other insects.
Insect gaits are usually obtained by two approaches: the centralized and the decentralized control architectures. Centralized controllers directly specify transitions of all legs, whereas in decentralized architectures, six nodes (legs) are connected in a parallel network; gaits arise by the interaction between neighbouring legs.
Foot coordination
The term "foot coordination" refers to the mechanism responsible for controlling the transition between footsteps; considering that the body does not turn around. Most approaches try to replicate the appearance of known insects, for example the tripod or tetrapod shape. However, other approaches have been used to find stable paces; for example, by launching programs using genetic algorithms or by optimizing the energy of walking.
The walking patterns of insects are usually obtained by two approaches: centralized and decentralized control architectures. The centralized controllers directly specify the transitions of all the legs, whereas in the decentralized architectures, six nodes (legs) are connected in a parallel network; the paces are obtained thanks to the interaction between the neighboring legs.
Controller of the paw
There are no borders to the complexity of the morphology of the paw. However, legs that are built on an insect model usually have between two and six degrees of freedom. Leg segments typically bear the name of their biological counterpart, which are similar for most species. From the body to the end of the paw, the segments bear the names of coxa, femur and tibia ; typically the joints between the coxa and the femur and between the femur and the tibia are considered as simple hinges. The models of the articulation between the body and the coxa comprise between one and three degrees of freedom, depending on the species and the thoracic segment on which the leg is enclosed.
Source from Wikipedia
2018年11月13日星期二
Legged robot
Legged robots are a type of mobile robot which use mechanical limbs for movement. They are more versatile than wheeled robots and can traverse many different terrains, though these advantages require increased complexity and power consumption. Legged robots often imitate legged animals, such as humans or insects, in an example of biomimicry.
Types
Legged robots can be categorized by the number of limbs they use, which determines gaits available. Many-legged robots tend to be more stable, while fewer legs lends itself to greater maneuverability.
One-legged
One-legged, or pogo stick robots use a hopping motion for navigation. In the 1980s, Carnegie Mellon University developed a one-legged robot to study balance. Berkeley's SALTO is another example.
Two-legged
Bipedal or two-legged robots exhibit bipedal motion. As such, they face two primary problems:
stability control, which refers to a robot's balance, and
motion control, which refers to a robot's ability to move.
Stability control is particularly difficult for bipedal systems, which must maintain balance in the forward-backward direction even at rest. Some robots, especially toys, solve this problem with large feet, which provide greater stability while reducing mobility. Alternatively, more advanced systems use sensors such as accelerometers or gyroscopes to provide dynamic feedback in a fashion that approximates a human being's balance. Such sensors are also employed for motion control and walking. The complexity of these tasks lends itself to machine learning.
Simple bipedal motion can be approximated by a rolling polygon where the length of each side matches that of a single step. As the step length grows shorter, the number of sides increases and the motion approaches that of a circle. This connects bipedal motion to wheeled motion as a limit of stride length.
Two-legged robots include:
Boston Dynamics' Atlas
Toy robots such as QRIO and ASIMO.
NASA's Valkyrie robot, intended to aid humans on Mars.
The ping-pong playing TOPIO robot.
Four-legged
Quadrupedal or four-legged robots exhibit quadrupedal motion. They benefit from increased stability over bipedal robots, especially during movement. At slow speeds, a quadrupedal robot may move only one leg at a time, ensuring a stable tripod. Four-legged robots also benefit from a lower center of gravity than two-legged systems.
Four legged robots include:
The TITAN series, developed since the 1980s by the Hirose-Yoneda Laboratory.
The dynamically stable BigDog, developed in 2005 by Boston Dynamics, NASA's Jet Propulsion Laboratory, and the Harvard University Concord Field Station.
BigDog's successor, the LS3.
Six-legged
Six-legged robots, or hexapods, are motivated by a desire for even greater stability than bipedal or quadrupedal robots. Their final designs often mimic the mechanics of insects, and their gaits may be categorized similarly. These include:
Wave gait: the slowest gait, in which pairs of legs move in a "wave" from the back to the front.
Tripod gait: a slightly faster step, in which three legs move at once. The remaining three legs provide a stable tripod for the robot.
Six-legged robots include:
Odex, a 375-pound hexapod developed by Odetics in the 1980s. Odex distinguished itself with its onboard computers, which controlled each leg.
Genghis, one of the earliest autonomous six-legged robots, was developed at MIT by Rodney Brooks in the 1980s.
The modern toy series, Hexbug.
Eight-legged
Eight-legged legged robots are inspired by spiders and other arachnids, as well as some underwater walkers. They offer by far the greatest stability, which enabled some early successes with legged robots.
Eight-legged robots include:
Dante, a Carnegie Mellon University project designed to explore Mount Erebus.
The T8X, a commercially available robot designed to emulate a spider's appearance and movements.
Hybrids
Some robots use a combination of legs and wheels. This grants a machine the speed and energy efficiency of wheeled locomotion as well as the mobility of legged navigation. Boston Dynamics' Handle, a bipedal robot with wheels on both legs, is one example.
Walking behavior
Static walking
Static walking is when the center of gravity of a robot is above the feet at all times, so that it can not fall over without the action of an external force.
Dynamic walking and running
Dynamic walking is when the center of gravity of a robot can also be outside the area of the feet without the robot falling down. In fact, one could speak of a "controlled fall" as the robot would fall in a sudden stop of its movement.
Dynamic walking is when the movement necessary to maintain speed results in a moment when no leg of the robot is touching the ground.
Static mobile robots
The classic walking robot consists of actuators , sensors and a computer control . The "legs" are usually moved by servomotors so that a predetermined movement program is unwound.
Two-legged static walking robot
The robot ASIMO moves with a maximum speed of 6 km / h, with a size of 1.30 m and a weight of 52 kg and he requires a lot of electrical energy. A special ability of him is that he can climb stairs.
Six-legged walking robot
Six-legged constructions are an ideal basis for statically stable walking robots. They are therefore suitable for movement on uneven terrain. There are two gaits (sequence of leg movements):
Tripod-course
Tetrapod transition
The tripod gear has three legs on the ground at any one time (example: Indian stick insect , with 3 stance and 3 swinging legs).
The tetrapod gait always has four legs on the ground (4 legs, 2 swinging legs).
In the case of gaiters with six orthogonal legs, a differentiation is also made according to the principle of movement of the legs apart from the sequence of leg movements:
Followers (follow the leader) (eg tripod gear, tetrapod gear)
Circular walker
Web runner (weaving walker)
Six-legged creatures run as followers. One leg follows (in whatever order) the other. Machines can do more. In the circular runner, the three legs of the right side have a common axis of rotation - like the hands of a clock (corresponding to the left legs). The rearmost leg is swung in front of the foremost leg. But how is the last leg to pass the other two legs? It simply swings under the belly (the robot platform).
The web runner also performs a biologically impossible movement. In the Web runner, all six legs sit on a common vertical axis in the center of the platform. Each leg can wander completely around the whole body (a horizontal telescopic movement makes it possible). The legs move from their rearmost position to the frontmost position by running around the other two legs on the outside.
When walking on uneven terrain, it is crucial that the robot find a safe touchdown point within its step size (foothold selection area) without having to deviate too far from its main heading direction.
Dynamic walking robot
Passive dynamic runners
Running robots that can move without energy source are based on a toy invented 150 years ago. It just had to be triggered and then could run down a small slope alone. To do this, the toy rocks from right to left and swings the leg straight up a little bit forward. Then it rocks from left to right and the other leg swings forward.
With this construction, the toy can move energy-efficiently and serve as a starting model for technically advanced running robots. In the 1980s, Tad McGeer had used the principle of the pendulum for stabilizing the movements , which was realized in this toy . No longer a complex and slow control system in a computer entrained should bring the robot to work, but the structure of the musculoskeletal system should stabilize the running robot without additional action. If the construction of the simple toy is supplemented with a "hip" or "movable feet", then such walking robots only need energy when accelerating the moving masses and no longer as in earlier running robots also when braking.
Source from Wikipedia
Types
Legged robots can be categorized by the number of limbs they use, which determines gaits available. Many-legged robots tend to be more stable, while fewer legs lends itself to greater maneuverability.
One-legged
One-legged, or pogo stick robots use a hopping motion for navigation. In the 1980s, Carnegie Mellon University developed a one-legged robot to study balance. Berkeley's SALTO is another example.
Two-legged
Bipedal or two-legged robots exhibit bipedal motion. As such, they face two primary problems:
stability control, which refers to a robot's balance, and
motion control, which refers to a robot's ability to move.
Stability control is particularly difficult for bipedal systems, which must maintain balance in the forward-backward direction even at rest. Some robots, especially toys, solve this problem with large feet, which provide greater stability while reducing mobility. Alternatively, more advanced systems use sensors such as accelerometers or gyroscopes to provide dynamic feedback in a fashion that approximates a human being's balance. Such sensors are also employed for motion control and walking. The complexity of these tasks lends itself to machine learning.
Simple bipedal motion can be approximated by a rolling polygon where the length of each side matches that of a single step. As the step length grows shorter, the number of sides increases and the motion approaches that of a circle. This connects bipedal motion to wheeled motion as a limit of stride length.
Two-legged robots include:
Boston Dynamics' Atlas
Toy robots such as QRIO and ASIMO.
NASA's Valkyrie robot, intended to aid humans on Mars.
The ping-pong playing TOPIO robot.
Four-legged
Quadrupedal or four-legged robots exhibit quadrupedal motion. They benefit from increased stability over bipedal robots, especially during movement. At slow speeds, a quadrupedal robot may move only one leg at a time, ensuring a stable tripod. Four-legged robots also benefit from a lower center of gravity than two-legged systems.
Four legged robots include:
The TITAN series, developed since the 1980s by the Hirose-Yoneda Laboratory.
The dynamically stable BigDog, developed in 2005 by Boston Dynamics, NASA's Jet Propulsion Laboratory, and the Harvard University Concord Field Station.
BigDog's successor, the LS3.
Six-legged
Six-legged robots, or hexapods, are motivated by a desire for even greater stability than bipedal or quadrupedal robots. Their final designs often mimic the mechanics of insects, and their gaits may be categorized similarly. These include:
Wave gait: the slowest gait, in which pairs of legs move in a "wave" from the back to the front.
Tripod gait: a slightly faster step, in which three legs move at once. The remaining three legs provide a stable tripod for the robot.
Six-legged robots include:
Odex, a 375-pound hexapod developed by Odetics in the 1980s. Odex distinguished itself with its onboard computers, which controlled each leg.
Genghis, one of the earliest autonomous six-legged robots, was developed at MIT by Rodney Brooks in the 1980s.
The modern toy series, Hexbug.
Eight-legged
Eight-legged legged robots are inspired by spiders and other arachnids, as well as some underwater walkers. They offer by far the greatest stability, which enabled some early successes with legged robots.
Eight-legged robots include:
Dante, a Carnegie Mellon University project designed to explore Mount Erebus.
The T8X, a commercially available robot designed to emulate a spider's appearance and movements.
Hybrids
Some robots use a combination of legs and wheels. This grants a machine the speed and energy efficiency of wheeled locomotion as well as the mobility of legged navigation. Boston Dynamics' Handle, a bipedal robot with wheels on both legs, is one example.
Walking behavior
Static walking
Static walking is when the center of gravity of a robot is above the feet at all times, so that it can not fall over without the action of an external force.
Dynamic walking and running
Dynamic walking is when the center of gravity of a robot can also be outside the area of the feet without the robot falling down. In fact, one could speak of a "controlled fall" as the robot would fall in a sudden stop of its movement.
Dynamic walking is when the movement necessary to maintain speed results in a moment when no leg of the robot is touching the ground.
Static mobile robots
The classic walking robot consists of actuators , sensors and a computer control . The "legs" are usually moved by servomotors so that a predetermined movement program is unwound.
Two-legged static walking robot
The robot ASIMO moves with a maximum speed of 6 km / h, with a size of 1.30 m and a weight of 52 kg and he requires a lot of electrical energy. A special ability of him is that he can climb stairs.
Six-legged walking robot
Six-legged constructions are an ideal basis for statically stable walking robots. They are therefore suitable for movement on uneven terrain. There are two gaits (sequence of leg movements):
Tripod-course
Tetrapod transition
The tripod gear has three legs on the ground at any one time (example: Indian stick insect , with 3 stance and 3 swinging legs).
The tetrapod gait always has four legs on the ground (4 legs, 2 swinging legs).
In the case of gaiters with six orthogonal legs, a differentiation is also made according to the principle of movement of the legs apart from the sequence of leg movements:
Followers (follow the leader) (eg tripod gear, tetrapod gear)
Circular walker
Web runner (weaving walker)
Six-legged creatures run as followers. One leg follows (in whatever order) the other. Machines can do more. In the circular runner, the three legs of the right side have a common axis of rotation - like the hands of a clock (corresponding to the left legs). The rearmost leg is swung in front of the foremost leg. But how is the last leg to pass the other two legs? It simply swings under the belly (the robot platform).
The web runner also performs a biologically impossible movement. In the Web runner, all six legs sit on a common vertical axis in the center of the platform. Each leg can wander completely around the whole body (a horizontal telescopic movement makes it possible). The legs move from their rearmost position to the frontmost position by running around the other two legs on the outside.
When walking on uneven terrain, it is crucial that the robot find a safe touchdown point within its step size (foothold selection area) without having to deviate too far from its main heading direction.
Dynamic walking robot
Passive dynamic runners
Running robots that can move without energy source are based on a toy invented 150 years ago. It just had to be triggered and then could run down a small slope alone. To do this, the toy rocks from right to left and swings the leg straight up a little bit forward. Then it rocks from left to right and the other leg swings forward.
With this construction, the toy can move energy-efficiently and serve as a starting model for technically advanced running robots. In the 1980s, Tad McGeer had used the principle of the pendulum for stabilizing the movements , which was realized in this toy . No longer a complex and slow control system in a computer entrained should bring the robot to work, but the structure of the musculoskeletal system should stabilize the running robot without additional action. If the construction of the simple toy is supplemented with a "hip" or "movable feet", then such walking robots only need energy when accelerating the moving masses and no longer as in earlier running robots also when braking.
Source from Wikipedia
Continuous track
Continuous track, also called tank tread or caterpillar track, is a system of vehicle propulsion in which a continuous band of treads or track plates is driven by two or more wheels. This band is typically made of modular steel plates in the case of military vehicles and heavy equipment, or synthetic rubber reinforced with steel wires in the case of lighter agricultural or construction vehicles.
The large surface area of the tracks distributes the weight of the vehicle better than steel or rubber tyres on an equivalent vehicle, enabling a continuous tracked vehicle to traverse soft ground with less likelihood of becoming stuck due to sinking. The prominent treads of the metal plates are both hard-wearing and damage resistant, especially in comparison to rubber tyres. The aggressive treads of the tracks provide good traction in soft surfaces but can damage paved surfaces, so some metal tracks can have rubber pads installed for use on paved surfaces.
Continuous tracks can be traced back as far as 1770 and today are commonly used on a variety of vehicles including bulldozers, excavators, tanks, and tractors, but can be found on any vehicle used in an application that can benefit from the added traction, low ground pressure and durability inherent in continuous track propulsion systems.
History
Polish mathematician and inventor Józef Maria Hoene-Wroński conceived of the idea in the 1830s. The British polymath Sir George Cayley patented a continuous track, which he called a "universal railway". In 1837, a Russian inventor Dmitry Zagryazhsky designed a "carriage with mobile tracks" which he patented the same year, but due to a lack of funds and interest from manufacturers he was unable to build a working prototype, and his patent was voided in 1839.
Dreadnaught wheel
Although not a continuous track in the form encountered today, a dreadnaught wheel or "endless railway wheel" was patented by the British Engineer James Boydell in 1846. In Boydell's design, a series of flat feet are attached to the periphery of the wheel, spreading the weight. A number of horse-drawn wagons, carts and gun carriages were successfully deployed in the Crimean War, waged between October 1853 and February 1856, the Royal Arsenal at Woolwich manufacturing dreadnaught wheels. A letter of recommendation was signed by Sir William Codrington, the General commanding the troops at Sebastopol.
Boydell patented improvements to his wheel in 1854 (No. 431) – the year his dreadnaught wheel was first applied to a steam engine – and 1858 (No. 356), the latter an impracticable palliative measure involving the lifting one or other of the driving wheels to facilitate turning.
A number of manufacturers including Richard Bach, Richard Garrett & Sons, Charles Burrell & Sons and Clayton & Shuttleworth applied the Boydell patent under licence. The British military were interested in Boydell's invention from an early date. One of the objectives was to transport Mallet's Mortar, a giant 36 in weapon which was under development, but, by the end of the Crimean war, the mortar was not ready for service. A detailed report of the tests on steam traction, carried out by a select Committee of the Board of Ordnance, was published in June 1856, by which date the Crimean War was over, consequently the mortar and its transportation became irrelevant. In those tests, a Garrett engine was put through its paces on Plumstead Common. The Garrett engine featured in the Lord Mayor's show in London, and in the following month that engine was shipped to Australia. A steam tractor employing dreadnaught wheels was built at Bach's Birmingham works, and was used between 1856 and 1858 for ploughing in Thetford; and the first generation of Burrell/Boydell engines was built at the St. Nicholas works in 1856, again, after the close of the Crimean war. Between late 1856 and 1862 Burrell manufactured not less than a score of engines fitted with dreadnaught wheels. In April 1858, "The Engineer" gave a brief description of a Clayton & Shuttleworth engine fitted with dreadnaught wheels, which was supplied not to the Western Allies, but to the Russian government for heavy artillery haulage in the Crimea, in the post-war period. Steam tractors fitted with dreadnaught wheels had a number of shortcomings and, notwithstanding the creations of the late 1850s, were never used extensively.
In August 1858, more than two years after the end of the Crimean War, John Fowler filed British Patent No. 1948 on another form of "Endless Railway". In his illustration of the invention, Fowler used a pair of wheels of equal diameter on each side of his vehicle, around which pair of toothed wheels ran a 'track' of eight jointed segments, with a smaller jockey/drive wheel between each pair of wheels, to support the 'track'. Comprising only eight sections, the 'track' sections are essentially 'longitudinal', as in Boydell's initial design. Fowler's arrangement is a precursor to the multi-section caterpillar track in which a relatively large number of short 'transverse' treads are used, as proposed by Sir George Caley in 1825, rather than a small number of relatively long 'longitudinal' treads.
Further to Fowler's patent of 1858, in 1877, a Russian, Fyodor Blinov, created a tracked vehicle called "wagon moved on endless rails" (caterpillars). It lacked self-propulsion and was pulled by horses. Blinov received a patent for his "wagon" in 1878. From 1881 to 1888 he developed a steam-powered caterpillar-tractor. This self-propelled crawler was successfully tested and featured at a farmers' exhibition in 1896.
Steam traction engines were used at the end of the 19th century in the Boer Wars. But neither dreadnaught wheels nor continuous tracks were used, rather "roll-out" wooden plank roads were thrown under the wheels as required.
In short, whilst the development of the continuous track engaged the attention of a number of inventors in the 18th and 19th centuries, the general use and exploitation of the continuous track belonged to the 20th century.
20th century
A little-known American inventor, Henry T. Stith, developed a continuous track prototype which was, in multiple forms, patented in 1873, 1880, and 1900. The last was for the application of the track to a prototype off-road bicycle built for his son. The 1900 prototype is retained by his surviving family.
Frank Beamond, a less-commonly known but significant British inventor, designed and built caterpillar tracks, and was granted patents for them in a number of countries, in 1900 and 1907.
Commercial success
An effective continuous track was invented and implemented by Alvin Orlando Lombard for the Lombard Steam Log Hauler. He was granted a patent in 1901 and built the first steam-powered log hauler at the Waterville Iron Works in Waterville, Maine, the same year. In all, 83 Lombard steam log haulers are known to have been built up to 1917, when production switched entirely to internal combustion engine powered machines, ending with a Fairbanks diesel-powered unit in 1934. Undoubtedly, Alvin Lombard was the first commercial manufacturer of the tractor crawler.
At least one of Lombard's steam-powered machines apparently remains in working order. A gasoline-powered Lombard hauler is on display at the Maine State Museum in Augusta. In addition, there may have been up to twice as many Phoenix Centipeed versions of the steam log hauler built under license from Lombard, with vertical instead of horizontal cylinders. In 1903, the founder of Holt Manufacturing, Benjamin Holt, paid Lombard $60,000 for the right to produce vehicles under his patent.
At about the same time a British agricultural company, Hornsby in Grantham, developed a continuous track which was patented in 1905. The design differed from modern tracks in that it flexed in only one direction, with the effect that the links locked together to form a solid rail on which the road wheels ran. Hornsby's tracked vehicles were given trials as artillery tractors by the British Army on several occasions between 1905 and 1910, but not adopted. The Hornsby tractors featured a track-steer clutch arrangement, which is the basis of the modern crawler operation. The patent was purchased by Holt.
Military application
Continuous track was first applied to a military vehicle on the British prototype tank Little Willie. British Army officers, Colonel Ernest Swinton and Colonel Maurice Hankey, became convinced that it was possible to develop a fighting vehicle that could provide protection from machine gun fire.
Caterpillar
The name came from a soldier during the tests on the Hornsby crawler, "trials began at Aldershot in July 1907. The soldiers immediately christened the 70bhp No.2 machine the 'caterpillar'."
Holt adopted that name for his "crawler" tractors. Holt began moving from steam to gasoline-powered designs, and in 1908 brought out the 40 horsepower "Holt Model 40 Caterpillar". Holt incorporated the Holt Caterpillar Company, in early 1910, later that year trademarked the name "Caterpillar" for his continuous tracks.
In a memorandum of 1908, Antarctic explorer Robert Falcon Scott presented his view that man-hauling to the South Pole was impossible and that motor traction was needed. Snow vehicles did not yet exist however, and so his engineer Reginald Skelton developed the idea of a caterpillar track for snow surfaces. These tracked motors were built by the Wolseley Tool and Motor Car Company in Birmingham, tested in Switzerland and Norway, and can be seen in action in Herbert Ponting's 1911 documentary film of Scott's Antarctic Terra Nova Expedition. Scott died during the expedition in 1912, but expedition member and biographer Apsley Cherry-Garrard credited Scott's "motors" with the inspiration for the British World War I tanks, writing: "Scott never knew their true possibilities; for they were the direct ancestors of the 'tanks' in France."
During World War I, Holt tractors were used by the British and Austro-Hungarian armies to tow heavy artillery and stimulated the development of tanks in several countries. The first tanks to go into action, the Mark I, built by Great Britain, were designed from scratch and were inspired by, but not directly based on, the Holt. The slightly later French and German tanks were built on modified Holt running gear.
Caterpillar Tractor Company began in 1925 from a merger of the Holt Manufacturing Company and the C. L. Best Tractor Company, an early successful manufacturer of crawler tractors.
With the Caterpillar D10 in 1977, Caterpillar resurrected a design by Holt and Best, the high-sprocket-drive, since known as the "High Drive", which had the advantage of keeping the main drive shaft away from ground shocks and dirt, and is still used in their larger dozers.
Patent history
A long line of patents disputes who was the "originator" of continuous tracks. There were a number of designs that attempted to achieve a track laying mechanism, although these designs do not generally resemble modern tracked vehicles.
Blinov
In 1877 Russian inventor Fyodor Abramovich Blinov created a tracked vehicle called "wagon moved on endless rails" (caterpillars). It lacked self-propelling and was horse-drawn. Blinov got a patent for his "wagon" the next year. Later, in 1881-1888 he created a steam-powered caterpillar-tractor. This self-propelled crawler was successfully tested and showed at a farmers' exhibition in 1896.
Dinsmoor
According to Scientific American, it was Charles Dinsmoor of Warren, Pennsylvania that invented a "vehicle" that was on endless tracks. The article gives a detailed description of the endless tracks and the illustration looks much like today's tracked vehicles. The invention has been patented as No. 351,749 on November 2, 1886.
Lombard
Alvin O. Lombard of Waterville, Maine was issued a patent in 1901 for the Lombard Steam Log Hauler that resembles a regular railroad steam locomotive with sled steerage on front and crawlers in rear for hauling logs in the Northeastern United States and Canada. The haulers allowed pulp to be taken to rivers in the winter. Prior to then, horses could be used only until snow depths made hauling impossible. Lombard began commercial production which lasted until around 1917 when focus switched entirely to gasoline powered machines. A gasoline-powered hauler is on display at the Maine State Museum in Augusta, Maine.
Hornsby/Holt/Phoenix
After Lombard began operations, Hornsby in England manufactured at least two full length "track steer" machines, and their patent was later purchased by Holt in 1913, allowing Holt to claim to be the "inventor" of the crawler tractor. Since the "tank" was a British concept it is more likely the Hornsby, which had been built and unsuccessfully pitched to their military, was the inspiration.
In a patent dispute involving rival crawler builder Best, testimony was brought in from people including Lombard, that Holt had inspected a Lombard log hauler shipped out to a western state by people who would later build the Phoenix log hauler in Eau Claire, Wisconsin, under license from Lombard. The Phoenix Centipeed typically had a fancier wood cab, steering wheel tipped forward at a 45 degree angle and vertical instead of horizontal cylinders.
Linn
In the meantime, a gasoline-powered motor home was built by Lombard for Holman Harry (Flannery) Linn of Old Town, Maine to pull the equipment wagon of his dog & pony show, resembling a trolley car only with wheels in front and Lombard crawlers in rear. Linn had experimented with gasoline and steam-powered vehicles and six-wheel drive before this, and at some point entered Lombard's employment as a demonstrator, mechanic and sales agent. This resulted in a question of proprietorship of patent rights after a single rear-tracked gasoline-powered road engine of tricycle arrangement was built to replace the larger motor home in 1909 on account of problems with the old picturesque wooden bridges. This dispute resulted in Linn departing Maine and relocating to Morris, New York, to build an improved, contour following flexible lag tread or crawler with independent suspension of halftrack type, gasoline and later diesel powered. Although several were delivered for military use between 1917 and 1946, Linn never received any large military orders. Most of the production between 1917 and 1952, approximately 2500 units, was sold directly to highway departments and contractors. Steel tracks and payload capacity allowed these machines to work in terrain that would typically cause the poorer quality rubber tyres that existed before the mid-1930s to spin uselessly, or shred completely.
Linn was a pioneer in snow removal before the practice was embraced in rural areas, with a nine-foot steel v-plow and sixteen foot adjustable leveling wings on either side. Once the highway system became paved, snowplowing could be done by four wheel drive trucks equipped by improving tyre designs, and the Linn became an off highway vehicle, for logging, mining, dam construction, arctic exploration, etc.
Engineering
Steering
Continuous track vehicle steer by applying more or less drive torque to one side of the vehicle than the other, and this can be implemented in a variety of ways.
Construction and operation
Modern tracks are built from modular chain links which together compose a closed chain. The links are jointed by a hinge, which allows the track to be flexible and wrap around a set of wheels to make an endless loop. The chain links are often broad, and can be made of manganese alloy steel for high strength, hardness, and abrasion resistance.
Track construction and assembly is dictated by the application. Military vehicles use a track shoe that is integral to the structure of the chain in order to reduce track weight. Reduced weight allows the vehicle to move faster and decreases overall vehicle weight to ease transportation. Since track weight is completely unsprung, reducing it improves suspension performance at speeds where the track's momentum is significant. In contrast, agricultural and construction vehicles opt for a track with shoes that attach to the chain with bolts and do not form part of the chain's structure. This allows track shoes to break without compromising the ability of the vehicle to move and decrease productivity but increases the overall weight of the track and vehicle.
The vehicle's weight is transferred to the bottom length of track by a number of road wheels, or sets of wheels called bogies. Road wheels are typically mounted on some form of suspension to cushion the ride over rough ground. Suspension design in military vehicles is a major area of development; the very early designs were often completely unsprung. Later-developed road wheel suspension offered only a few inches of travel using springs, whereas modern hydro-pneumatic systems allow several feet of travel and include shock absorbers. Torsion-bar suspension has become the most common type of military vehicle suspension. Construction vehicles have smaller road wheels that are designed primarily to prevent track derailment and they are normally contained in a single bogie that includes the idler-wheel and sometimes the sprocket.
Transfer of power to the track is accomplished by a drive wheel, or drive sprocket, driven by the motor and engaging with holes in the track links or with pegs on them to drive the track. In military vehicles, the drive wheel is typically mounted well above the contact area on the ground, allowing it to be fixed in position. In agricultural crawlers it is normally incorporated as part of the bogie. Placing suspension on the sprocket is possible, but is mechanically more complicated. A non-powered wheel, an idler, is placed at the opposite end of the track, primarily to tension the track, since loose track could be easily thrown (slipped) off the wheels. To prevent throwing, the inner surface of the track links usually have vertical guide horns engaging grooves, or gaps between the doubled road and idler/sprocket wheels. In military vehicles with a rear sprocket, the idler wheel is placed higher than the road wheels to allow it to climb over obstacles. Some track arrangements use return rollers to keep the top of the track running straight between the drive sprocket and idler. Others, called slack track, allow the track to droop and run along the tops of large road wheels. This was a feature of the Christie suspension, leading to occasional misidentification of other slack track-equipped vehicles.
Overlapping road wheels
Many World War II German military vehicles, initially (starting in the late 1930s) including all vehicles originally designed to be half-tracks and all later tank designs (after the Panzer IV), had slack-track systems, usually driven by a front-located drive sprocket, the track returning along the tops of a design of overlapping and sometimes interleaved large diameter road wheels, as on the suspension systems of the Tiger I and Panther tanks, generically known by the term Schachtellaufwerk in German, for both half-track and fully tracked vehicles. There were suspensions with one (sometimes double) wheel per axle, alternately supporting the inner and outer side of the track, and interleaved suspensions with two or three road wheels per axle, distributing the load over the track.
The choice of overlapping/interleaved road wheels allowed the use of slightly more torsion bar suspension members, allowing any German tracked military vehicle with such a setup to have a noticeably smoother ride over challenging terrain, leading to reduced wear and more accurate fire. There were some major disadvantages with this though, one being on the Russian front, mud would get stuck in-between the overlapping wheels, and would freeze, immobilizing the vehicle. As a tracked vehicle moves, the load of each wheel moves over the track, pushing down and forward that part of the earth, snow, etc. under it, similarly to a wheeled vehicle but to a lesser extent because the tread helps distribute the load. Apparently, on some surfaces, this consumes enough energy to slow the vehicle down significantly, so overlapped and interleaved wheels improve performance (including fuel consumption) by loading the track more evenly. It also must have extended the life of the tracks and possibly of the wheels. The wheels also better protect the vehicle from enemy fire, and mobility is improved when some wheels are missing. But this complicated approach has not been used since World War II ended. This may be related more to maintenance than to original cost. The torsion bars and bearings may stay dry and clean, but the wheels and tread work in mud, sand, rocks, snow and so on. In addition, the outer wheels (up to 9 of them, some double) had to be removed to access the inner ones. In WW II, vehicles typically had to be maintained a few months before being destroyed or captured, but in peacetime vehicles must train several crews, over a period of decades.
Advantages
Tracked vehicles have better mobility over rough terrain than those with wheels. They smooth out the bumps, glide over small obstacles and are capable of crossing trenches or breaks in the terrain. Riding in a fast tracked vehicle feels like riding in a boat over heavy swells. Tracks cannot be punctured or torn. Tracks are much less likely to get stuck in soft ground, mud, or snow since they distribute the weight of the vehicle over a larger contact area, decreasing its ground pressure. In addition, the larger contact area, coupled with the cleats, or grousers, on the track shoes, allows vastly superior traction that results in a much better ability to push or pull large loads where wheeled vehicles would dig in. Bulldozers, which are most often tracked, use this attribute to rescue other vehicles, (such as wheel loaders) which have become stuck in, or sunk into, the ground. Tracks can also give higher maneuverability, as tracked vehicles can turn in place without forward or backward movement by driving the tracks in opposite directions. In addition, should a track be broken, assuming the correct tools are available, it can be repaired without the need for special facilities; something which is crucial in a combat situation.
The seventy-ton M1 Abrams tank has an average ground pressure of just over 15 psi (100 kPa). Since tyre air pressure is approximately equal to average ground pressure, a typical car will have an average ground pressure of 28 psi (190 kPa) to 33 psi (230 kPa).
Disadvantages
The disadvantages of tracks are lower top speed, much greater mechanical complexity, shorter life and the damage that their all-steel versions cause to the surface on which they pass. They are assumed to severely damage hard terrain such as asphalt pavement, but actually have significantly lower ground pressures than equivalent or lighter wheeled vehicles. However, they often cause damage to less firm terrain such as lawns, gravel roads, and farm fields, as the sharp edges of the track easily rout the turf. Accordingly, vehicle laws and local ordinances often require rubberised tracks or track pads. A compromise between all-steel and all-rubber tracks exists: attaching rubber pads to individual track links ensures that continuous track vehicles can travel more smoothly, quickly, and quietly on paved surfaces. While these pads slightly reduce a vehicle's cross-country traction, in theory they prevent damage to any pavement.
Additionally, the loss of a single segment in a track immobilizes the entire vehicle, which can be a disadvantage in situations where high reliability is important. Tracks can also ride off their guide wheels, idlers or sprockets, which can cause them to jam or to come completely off the guide system (this is called a 'thrown' track). Jammed tracks may become so tight that the track may need to be broken before a repair is possible, which requires either explosives or special tools. Multi-wheeled vehicles, for example, 8 X 8 military vehicles, may often continue driving even after the loss of one or more non-sequential wheels, depending on the base wheel pattern and drive train.
Many manufacturers provide rubber tracks instead of steel, especially for agricultural applications. Rather than a track made of linked steel plates, a reinforced rubber belt with chevron treads is used. In comparison to steel tracks, rubber tracks are lighter, make less noise, create less maximal ground pressure and do not damage paved roads. The disadvantage is that they are not as solid as steel tracks. Previous belt-like systems, such as those used for half-tracks in World War II, were not as strong, and during military actions were easily damaged. The first rubber track was invented and constructed by Adolphe Kégresse and patented in 1913; rubber tracks are often called Kégresse tracks.
Prolonged use places enormous strain on the drive transmission and the mechanics of the tracks, which must be overhauled or replaced regularly. It is common to see tracked vehicles such as bulldozers or tanks transported long distances by a wheeled carrier such as a tank transporter or train, though technological advances have made this practice less common among tracked military vehicles than it once was.
"Live" and "dead" track
Tracks may be broadly categorized as "live" or "dead" track. "Dead" track is a simple design in which each track plate is connected to the rest with hinge-type pins. These dead tracks will lie flat if placed on the ground; the drive sprocket pulls the track around the wheels with no assistance from the track itself. "Live" track is slightly more complex, with each link connected to the next by a bushing which causes the track to bend slightly inward. A length of live track left on the ground will curl upward slightly at each end. Although the drive sprocket must still pull the track around the wheels, the track itself tends to bend inward, slightly assisting the sprocket and somewhat conforming to the wheels.
Current manufacturers
The pioneer manufacturers have been replaced mostly by large tractor companies such as AGCO, Liebherr Group, John Deere, Yanmar, New Holland, Kubota, Case, Caterpillar Inc., CLAAS. Also, there are some crawler tractor companies specialising in niche markets. Examples are Otter Mfg. Co. and Struck Corporation., with many wheeled vehicle conversion kits available from the American Mattracks firm of Minnesota since the mid-1990s.
Russian off-road vehicles are built by companies such as ZZGT and Vityaz.
In nature
Navicula diatoms are known for their ability to creep about on each other and on hard surfaces such as microscope slides. It is thought that around the outside of the navicula's shell is a girdle of protoplasm that can flow and thus act as a tank track.
Source from Wikipedia
The large surface area of the tracks distributes the weight of the vehicle better than steel or rubber tyres on an equivalent vehicle, enabling a continuous tracked vehicle to traverse soft ground with less likelihood of becoming stuck due to sinking. The prominent treads of the metal plates are both hard-wearing and damage resistant, especially in comparison to rubber tyres. The aggressive treads of the tracks provide good traction in soft surfaces but can damage paved surfaces, so some metal tracks can have rubber pads installed for use on paved surfaces.
Continuous tracks can be traced back as far as 1770 and today are commonly used on a variety of vehicles including bulldozers, excavators, tanks, and tractors, but can be found on any vehicle used in an application that can benefit from the added traction, low ground pressure and durability inherent in continuous track propulsion systems.
History
Polish mathematician and inventor Józef Maria Hoene-Wroński conceived of the idea in the 1830s. The British polymath Sir George Cayley patented a continuous track, which he called a "universal railway". In 1837, a Russian inventor Dmitry Zagryazhsky designed a "carriage with mobile tracks" which he patented the same year, but due to a lack of funds and interest from manufacturers he was unable to build a working prototype, and his patent was voided in 1839.
Dreadnaught wheel
Although not a continuous track in the form encountered today, a dreadnaught wheel or "endless railway wheel" was patented by the British Engineer James Boydell in 1846. In Boydell's design, a series of flat feet are attached to the periphery of the wheel, spreading the weight. A number of horse-drawn wagons, carts and gun carriages were successfully deployed in the Crimean War, waged between October 1853 and February 1856, the Royal Arsenal at Woolwich manufacturing dreadnaught wheels. A letter of recommendation was signed by Sir William Codrington, the General commanding the troops at Sebastopol.
Boydell patented improvements to his wheel in 1854 (No. 431) – the year his dreadnaught wheel was first applied to a steam engine – and 1858 (No. 356), the latter an impracticable palliative measure involving the lifting one or other of the driving wheels to facilitate turning.
A number of manufacturers including Richard Bach, Richard Garrett & Sons, Charles Burrell & Sons and Clayton & Shuttleworth applied the Boydell patent under licence. The British military were interested in Boydell's invention from an early date. One of the objectives was to transport Mallet's Mortar, a giant 36 in weapon which was under development, but, by the end of the Crimean war, the mortar was not ready for service. A detailed report of the tests on steam traction, carried out by a select Committee of the Board of Ordnance, was published in June 1856, by which date the Crimean War was over, consequently the mortar and its transportation became irrelevant. In those tests, a Garrett engine was put through its paces on Plumstead Common. The Garrett engine featured in the Lord Mayor's show in London, and in the following month that engine was shipped to Australia. A steam tractor employing dreadnaught wheels was built at Bach's Birmingham works, and was used between 1856 and 1858 for ploughing in Thetford; and the first generation of Burrell/Boydell engines was built at the St. Nicholas works in 1856, again, after the close of the Crimean war. Between late 1856 and 1862 Burrell manufactured not less than a score of engines fitted with dreadnaught wheels. In April 1858, "The Engineer" gave a brief description of a Clayton & Shuttleworth engine fitted with dreadnaught wheels, which was supplied not to the Western Allies, but to the Russian government for heavy artillery haulage in the Crimea, in the post-war period. Steam tractors fitted with dreadnaught wheels had a number of shortcomings and, notwithstanding the creations of the late 1850s, were never used extensively.
In August 1858, more than two years after the end of the Crimean War, John Fowler filed British Patent No. 1948 on another form of "Endless Railway". In his illustration of the invention, Fowler used a pair of wheels of equal diameter on each side of his vehicle, around which pair of toothed wheels ran a 'track' of eight jointed segments, with a smaller jockey/drive wheel between each pair of wheels, to support the 'track'. Comprising only eight sections, the 'track' sections are essentially 'longitudinal', as in Boydell's initial design. Fowler's arrangement is a precursor to the multi-section caterpillar track in which a relatively large number of short 'transverse' treads are used, as proposed by Sir George Caley in 1825, rather than a small number of relatively long 'longitudinal' treads.
Further to Fowler's patent of 1858, in 1877, a Russian, Fyodor Blinov, created a tracked vehicle called "wagon moved on endless rails" (caterpillars). It lacked self-propulsion and was pulled by horses. Blinov received a patent for his "wagon" in 1878. From 1881 to 1888 he developed a steam-powered caterpillar-tractor. This self-propelled crawler was successfully tested and featured at a farmers' exhibition in 1896.
Steam traction engines were used at the end of the 19th century in the Boer Wars. But neither dreadnaught wheels nor continuous tracks were used, rather "roll-out" wooden plank roads were thrown under the wheels as required.
In short, whilst the development of the continuous track engaged the attention of a number of inventors in the 18th and 19th centuries, the general use and exploitation of the continuous track belonged to the 20th century.
20th century
A little-known American inventor, Henry T. Stith, developed a continuous track prototype which was, in multiple forms, patented in 1873, 1880, and 1900. The last was for the application of the track to a prototype off-road bicycle built for his son. The 1900 prototype is retained by his surviving family.
Frank Beamond, a less-commonly known but significant British inventor, designed and built caterpillar tracks, and was granted patents for them in a number of countries, in 1900 and 1907.
Commercial success
An effective continuous track was invented and implemented by Alvin Orlando Lombard for the Lombard Steam Log Hauler. He was granted a patent in 1901 and built the first steam-powered log hauler at the Waterville Iron Works in Waterville, Maine, the same year. In all, 83 Lombard steam log haulers are known to have been built up to 1917, when production switched entirely to internal combustion engine powered machines, ending with a Fairbanks diesel-powered unit in 1934. Undoubtedly, Alvin Lombard was the first commercial manufacturer of the tractor crawler.
At least one of Lombard's steam-powered machines apparently remains in working order. A gasoline-powered Lombard hauler is on display at the Maine State Museum in Augusta. In addition, there may have been up to twice as many Phoenix Centipeed versions of the steam log hauler built under license from Lombard, with vertical instead of horizontal cylinders. In 1903, the founder of Holt Manufacturing, Benjamin Holt, paid Lombard $60,000 for the right to produce vehicles under his patent.
At about the same time a British agricultural company, Hornsby in Grantham, developed a continuous track which was patented in 1905. The design differed from modern tracks in that it flexed in only one direction, with the effect that the links locked together to form a solid rail on which the road wheels ran. Hornsby's tracked vehicles were given trials as artillery tractors by the British Army on several occasions between 1905 and 1910, but not adopted. The Hornsby tractors featured a track-steer clutch arrangement, which is the basis of the modern crawler operation. The patent was purchased by Holt.
Military application
Continuous track was first applied to a military vehicle on the British prototype tank Little Willie. British Army officers, Colonel Ernest Swinton and Colonel Maurice Hankey, became convinced that it was possible to develop a fighting vehicle that could provide protection from machine gun fire.
Caterpillar
The name came from a soldier during the tests on the Hornsby crawler, "trials began at Aldershot in July 1907. The soldiers immediately christened the 70bhp No.2 machine the 'caterpillar'."
Holt adopted that name for his "crawler" tractors. Holt began moving from steam to gasoline-powered designs, and in 1908 brought out the 40 horsepower "Holt Model 40 Caterpillar". Holt incorporated the Holt Caterpillar Company, in early 1910, later that year trademarked the name "Caterpillar" for his continuous tracks.
In a memorandum of 1908, Antarctic explorer Robert Falcon Scott presented his view that man-hauling to the South Pole was impossible and that motor traction was needed. Snow vehicles did not yet exist however, and so his engineer Reginald Skelton developed the idea of a caterpillar track for snow surfaces. These tracked motors were built by the Wolseley Tool and Motor Car Company in Birmingham, tested in Switzerland and Norway, and can be seen in action in Herbert Ponting's 1911 documentary film of Scott's Antarctic Terra Nova Expedition. Scott died during the expedition in 1912, but expedition member and biographer Apsley Cherry-Garrard credited Scott's "motors" with the inspiration for the British World War I tanks, writing: "Scott never knew their true possibilities; for they were the direct ancestors of the 'tanks' in France."
During World War I, Holt tractors were used by the British and Austro-Hungarian armies to tow heavy artillery and stimulated the development of tanks in several countries. The first tanks to go into action, the Mark I, built by Great Britain, were designed from scratch and were inspired by, but not directly based on, the Holt. The slightly later French and German tanks were built on modified Holt running gear.
Caterpillar Tractor Company began in 1925 from a merger of the Holt Manufacturing Company and the C. L. Best Tractor Company, an early successful manufacturer of crawler tractors.
With the Caterpillar D10 in 1977, Caterpillar resurrected a design by Holt and Best, the high-sprocket-drive, since known as the "High Drive", which had the advantage of keeping the main drive shaft away from ground shocks and dirt, and is still used in their larger dozers.
Patent history
A long line of patents disputes who was the "originator" of continuous tracks. There were a number of designs that attempted to achieve a track laying mechanism, although these designs do not generally resemble modern tracked vehicles.
Blinov
In 1877 Russian inventor Fyodor Abramovich Blinov created a tracked vehicle called "wagon moved on endless rails" (caterpillars). It lacked self-propelling and was horse-drawn. Blinov got a patent for his "wagon" the next year. Later, in 1881-1888 he created a steam-powered caterpillar-tractor. This self-propelled crawler was successfully tested and showed at a farmers' exhibition in 1896.
Dinsmoor
According to Scientific American, it was Charles Dinsmoor of Warren, Pennsylvania that invented a "vehicle" that was on endless tracks. The article gives a detailed description of the endless tracks and the illustration looks much like today's tracked vehicles. The invention has been patented as No. 351,749 on November 2, 1886.
Lombard
Alvin O. Lombard of Waterville, Maine was issued a patent in 1901 for the Lombard Steam Log Hauler that resembles a regular railroad steam locomotive with sled steerage on front and crawlers in rear for hauling logs in the Northeastern United States and Canada. The haulers allowed pulp to be taken to rivers in the winter. Prior to then, horses could be used only until snow depths made hauling impossible. Lombard began commercial production which lasted until around 1917 when focus switched entirely to gasoline powered machines. A gasoline-powered hauler is on display at the Maine State Museum in Augusta, Maine.
Hornsby/Holt/Phoenix
After Lombard began operations, Hornsby in England manufactured at least two full length "track steer" machines, and their patent was later purchased by Holt in 1913, allowing Holt to claim to be the "inventor" of the crawler tractor. Since the "tank" was a British concept it is more likely the Hornsby, which had been built and unsuccessfully pitched to their military, was the inspiration.
In a patent dispute involving rival crawler builder Best, testimony was brought in from people including Lombard, that Holt had inspected a Lombard log hauler shipped out to a western state by people who would later build the Phoenix log hauler in Eau Claire, Wisconsin, under license from Lombard. The Phoenix Centipeed typically had a fancier wood cab, steering wheel tipped forward at a 45 degree angle and vertical instead of horizontal cylinders.
Linn
In the meantime, a gasoline-powered motor home was built by Lombard for Holman Harry (Flannery) Linn of Old Town, Maine to pull the equipment wagon of his dog & pony show, resembling a trolley car only with wheels in front and Lombard crawlers in rear. Linn had experimented with gasoline and steam-powered vehicles and six-wheel drive before this, and at some point entered Lombard's employment as a demonstrator, mechanic and sales agent. This resulted in a question of proprietorship of patent rights after a single rear-tracked gasoline-powered road engine of tricycle arrangement was built to replace the larger motor home in 1909 on account of problems with the old picturesque wooden bridges. This dispute resulted in Linn departing Maine and relocating to Morris, New York, to build an improved, contour following flexible lag tread or crawler with independent suspension of halftrack type, gasoline and later diesel powered. Although several were delivered for military use between 1917 and 1946, Linn never received any large military orders. Most of the production between 1917 and 1952, approximately 2500 units, was sold directly to highway departments and contractors. Steel tracks and payload capacity allowed these machines to work in terrain that would typically cause the poorer quality rubber tyres that existed before the mid-1930s to spin uselessly, or shred completely.
Linn was a pioneer in snow removal before the practice was embraced in rural areas, with a nine-foot steel v-plow and sixteen foot adjustable leveling wings on either side. Once the highway system became paved, snowplowing could be done by four wheel drive trucks equipped by improving tyre designs, and the Linn became an off highway vehicle, for logging, mining, dam construction, arctic exploration, etc.
Engineering
Steering
Continuous track vehicle steer by applying more or less drive torque to one side of the vehicle than the other, and this can be implemented in a variety of ways.
Construction and operation
Modern tracks are built from modular chain links which together compose a closed chain. The links are jointed by a hinge, which allows the track to be flexible and wrap around a set of wheels to make an endless loop. The chain links are often broad, and can be made of manganese alloy steel for high strength, hardness, and abrasion resistance.
Track construction and assembly is dictated by the application. Military vehicles use a track shoe that is integral to the structure of the chain in order to reduce track weight. Reduced weight allows the vehicle to move faster and decreases overall vehicle weight to ease transportation. Since track weight is completely unsprung, reducing it improves suspension performance at speeds where the track's momentum is significant. In contrast, agricultural and construction vehicles opt for a track with shoes that attach to the chain with bolts and do not form part of the chain's structure. This allows track shoes to break without compromising the ability of the vehicle to move and decrease productivity but increases the overall weight of the track and vehicle.
The vehicle's weight is transferred to the bottom length of track by a number of road wheels, or sets of wheels called bogies. Road wheels are typically mounted on some form of suspension to cushion the ride over rough ground. Suspension design in military vehicles is a major area of development; the very early designs were often completely unsprung. Later-developed road wheel suspension offered only a few inches of travel using springs, whereas modern hydro-pneumatic systems allow several feet of travel and include shock absorbers. Torsion-bar suspension has become the most common type of military vehicle suspension. Construction vehicles have smaller road wheels that are designed primarily to prevent track derailment and they are normally contained in a single bogie that includes the idler-wheel and sometimes the sprocket.
Transfer of power to the track is accomplished by a drive wheel, or drive sprocket, driven by the motor and engaging with holes in the track links or with pegs on them to drive the track. In military vehicles, the drive wheel is typically mounted well above the contact area on the ground, allowing it to be fixed in position. In agricultural crawlers it is normally incorporated as part of the bogie. Placing suspension on the sprocket is possible, but is mechanically more complicated. A non-powered wheel, an idler, is placed at the opposite end of the track, primarily to tension the track, since loose track could be easily thrown (slipped) off the wheels. To prevent throwing, the inner surface of the track links usually have vertical guide horns engaging grooves, or gaps between the doubled road and idler/sprocket wheels. In military vehicles with a rear sprocket, the idler wheel is placed higher than the road wheels to allow it to climb over obstacles. Some track arrangements use return rollers to keep the top of the track running straight between the drive sprocket and idler. Others, called slack track, allow the track to droop and run along the tops of large road wheels. This was a feature of the Christie suspension, leading to occasional misidentification of other slack track-equipped vehicles.
Overlapping road wheels
Many World War II German military vehicles, initially (starting in the late 1930s) including all vehicles originally designed to be half-tracks and all later tank designs (after the Panzer IV), had slack-track systems, usually driven by a front-located drive sprocket, the track returning along the tops of a design of overlapping and sometimes interleaved large diameter road wheels, as on the suspension systems of the Tiger I and Panther tanks, generically known by the term Schachtellaufwerk in German, for both half-track and fully tracked vehicles. There were suspensions with one (sometimes double) wheel per axle, alternately supporting the inner and outer side of the track, and interleaved suspensions with two or three road wheels per axle, distributing the load over the track.
The choice of overlapping/interleaved road wheels allowed the use of slightly more torsion bar suspension members, allowing any German tracked military vehicle with such a setup to have a noticeably smoother ride over challenging terrain, leading to reduced wear and more accurate fire. There were some major disadvantages with this though, one being on the Russian front, mud would get stuck in-between the overlapping wheels, and would freeze, immobilizing the vehicle. As a tracked vehicle moves, the load of each wheel moves over the track, pushing down and forward that part of the earth, snow, etc. under it, similarly to a wheeled vehicle but to a lesser extent because the tread helps distribute the load. Apparently, on some surfaces, this consumes enough energy to slow the vehicle down significantly, so overlapped and interleaved wheels improve performance (including fuel consumption) by loading the track more evenly. It also must have extended the life of the tracks and possibly of the wheels. The wheels also better protect the vehicle from enemy fire, and mobility is improved when some wheels are missing. But this complicated approach has not been used since World War II ended. This may be related more to maintenance than to original cost. The torsion bars and bearings may stay dry and clean, but the wheels and tread work in mud, sand, rocks, snow and so on. In addition, the outer wheels (up to 9 of them, some double) had to be removed to access the inner ones. In WW II, vehicles typically had to be maintained a few months before being destroyed or captured, but in peacetime vehicles must train several crews, over a period of decades.
Advantages
Tracked vehicles have better mobility over rough terrain than those with wheels. They smooth out the bumps, glide over small obstacles and are capable of crossing trenches or breaks in the terrain. Riding in a fast tracked vehicle feels like riding in a boat over heavy swells. Tracks cannot be punctured or torn. Tracks are much less likely to get stuck in soft ground, mud, or snow since they distribute the weight of the vehicle over a larger contact area, decreasing its ground pressure. In addition, the larger contact area, coupled with the cleats, or grousers, on the track shoes, allows vastly superior traction that results in a much better ability to push or pull large loads where wheeled vehicles would dig in. Bulldozers, which are most often tracked, use this attribute to rescue other vehicles, (such as wheel loaders) which have become stuck in, or sunk into, the ground. Tracks can also give higher maneuverability, as tracked vehicles can turn in place without forward or backward movement by driving the tracks in opposite directions. In addition, should a track be broken, assuming the correct tools are available, it can be repaired without the need for special facilities; something which is crucial in a combat situation.
The seventy-ton M1 Abrams tank has an average ground pressure of just over 15 psi (100 kPa). Since tyre air pressure is approximately equal to average ground pressure, a typical car will have an average ground pressure of 28 psi (190 kPa) to 33 psi (230 kPa).
Disadvantages
The disadvantages of tracks are lower top speed, much greater mechanical complexity, shorter life and the damage that their all-steel versions cause to the surface on which they pass. They are assumed to severely damage hard terrain such as asphalt pavement, but actually have significantly lower ground pressures than equivalent or lighter wheeled vehicles. However, they often cause damage to less firm terrain such as lawns, gravel roads, and farm fields, as the sharp edges of the track easily rout the turf. Accordingly, vehicle laws and local ordinances often require rubberised tracks or track pads. A compromise between all-steel and all-rubber tracks exists: attaching rubber pads to individual track links ensures that continuous track vehicles can travel more smoothly, quickly, and quietly on paved surfaces. While these pads slightly reduce a vehicle's cross-country traction, in theory they prevent damage to any pavement.
Additionally, the loss of a single segment in a track immobilizes the entire vehicle, which can be a disadvantage in situations where high reliability is important. Tracks can also ride off their guide wheels, idlers or sprockets, which can cause them to jam or to come completely off the guide system (this is called a 'thrown' track). Jammed tracks may become so tight that the track may need to be broken before a repair is possible, which requires either explosives or special tools. Multi-wheeled vehicles, for example, 8 X 8 military vehicles, may often continue driving even after the loss of one or more non-sequential wheels, depending on the base wheel pattern and drive train.
Many manufacturers provide rubber tracks instead of steel, especially for agricultural applications. Rather than a track made of linked steel plates, a reinforced rubber belt with chevron treads is used. In comparison to steel tracks, rubber tracks are lighter, make less noise, create less maximal ground pressure and do not damage paved roads. The disadvantage is that they are not as solid as steel tracks. Previous belt-like systems, such as those used for half-tracks in World War II, were not as strong, and during military actions were easily damaged. The first rubber track was invented and constructed by Adolphe Kégresse and patented in 1913; rubber tracks are often called Kégresse tracks.
Prolonged use places enormous strain on the drive transmission and the mechanics of the tracks, which must be overhauled or replaced regularly. It is common to see tracked vehicles such as bulldozers or tanks transported long distances by a wheeled carrier such as a tank transporter or train, though technological advances have made this practice less common among tracked military vehicles than it once was.
"Live" and "dead" track
Tracks may be broadly categorized as "live" or "dead" track. "Dead" track is a simple design in which each track plate is connected to the rest with hinge-type pins. These dead tracks will lie flat if placed on the ground; the drive sprocket pulls the track around the wheels with no assistance from the track itself. "Live" track is slightly more complex, with each link connected to the next by a bushing which causes the track to bend slightly inward. A length of live track left on the ground will curl upward slightly at each end. Although the drive sprocket must still pull the track around the wheels, the track itself tends to bend inward, slightly assisting the sprocket and somewhat conforming to the wheels.
Current manufacturers
The pioneer manufacturers have been replaced mostly by large tractor companies such as AGCO, Liebherr Group, John Deere, Yanmar, New Holland, Kubota, Case, Caterpillar Inc., CLAAS. Also, there are some crawler tractor companies specialising in niche markets. Examples are Otter Mfg. Co. and Struck Corporation., with many wheeled vehicle conversion kits available from the American Mattracks firm of Minnesota since the mid-1990s.
Russian off-road vehicles are built by companies such as ZZGT and Vityaz.
In nature
Navicula diatoms are known for their ability to creep about on each other and on hard surfaces such as microscope slides. It is thought that around the outside of the navicula's shell is a girdle of protoplasm that can flow and thus act as a tank track.
Source from Wikipedia
Autonomous underwater vehicle
An autonomous underwater vehicle (AUV) is a robot that travels underwater without requiring input from an operator. AUVs constitute part of a larger group of undersea systems known as unmanned underwater vehicles, a classification that includes non-autonomous remotely operated underwater vehicles (ROVs) – controlled and powered from the surface by an operator/pilot via an umbilical or using remote control. In military applications a AUV is more often referred to as unmanned undersea vehicle (UUV). Underwater gliders are a subclass of AUVs.
History
The first AUV was developed at the Applied Physics Laboratory at the University of Washington as early as 1957 by Stan Murphy, Bob Francois and later on, Terry Ewart. The "Special Purpose Underwater Research Vehicle", or SPURV, was used to study diffusion, acoustic transmission, and submarine wakes.
Other early AUVs were developed at the Massachusetts Institute of Technology in the 1970s. One of these is on display in the Hart Nautical Gallery in MIT. At the same time, AUVs were also developed in the Soviet Union (although this was not commonly known until much later).
Applications
Until relatively recently, AUVs have been used for a limited number of tasks dictated by the technology available. With the development of more advanced processing capabilities and high yield power supplies, AUVs are now being used for more and more tasks with roles and missions constantly evolving.
Commercial
The oil and gas industry uses AUVs to make detailed maps of the seafloor before they start building subsea infrastructure; pipelines and sub sea completions can be installed in the most cost effective manner with minimum disruption to the environment. The AUV allows survey companies to conduct precise surveys of areas where traditional bathymetric surveys would be less effective or too costly. Also, post-lay pipe surveys are now possible, which includes pipeline inspection. The use of AUVs for pipeline inspection and inspection of underwater man-made structures is becoming more common.
Research
A University of South Florida researcher deploys Tavros02, a solar-powered "tweeting" AUV (SAUV)
Scientists use AUVs to study lakes, the ocean, and the ocean floor. A variety of sensors can be affixed to AUVs to measure the concentration of various elements or compounds, the absorption or reflection of light, and the presence of microscopic life. Examples include conductivity-temperature-depth sensors (CTDs), fluorometers, and pH sensors. Additionally, AUVs can be configured as tow-vehicles to deliver customized sensor packages to specific locations.
Hobby
Many roboticists construct AUVs as a hobby. Several competitions exist which allow these homemade AUVs to compete against each other while accomplishing objectives. Like their commercial brethren, these AUVs can be fitted with cameras, lights, or sonar. As a consequence of limited resources and inexperience, hobbyist AUVs can rarely compete with commercial models on operational depth, durability, or sophistication. Finally, these hobby AUVs are usually not oceangoing, being operated most of the time in pools or lake beds. A simple AUV can be constructed from a microcontroller, PVC pressure housing, automatic door lock actuator, syringes, and a DPDT relay. Some participants in competitions create open-source designs.
Illegal drug traffic
Submarines that travel autonomously to a destination by means of GPS navigation have been made by illegal drug traffickers.
Air crash investigations
Autonomous underwater vehicles, for example AUV ABYSS, have been used to find wreckages of missing airplanes, e.g. Air France Flight 447, and the Bluefin-21 AUV was used in the search for Malaysia Airlines Flight 370.
Military applications
The U.S. Navy Unmanned Undersea Vehicle (UUV) Master Plan identified the following UUV's missions:
Intelligence, surveillance, and reconnaissance
Mine countermeasures
Anti-submarine warfare
Inspection/identification
Oceanography
Communication/navigation network nodes
Payload delivery
Information operations
Time-critical strike
The Navy Master Plan divided all UUVs into four classes:
Man-portable vehicle class: 25–100 lb displacement; 10–20 hours endurance; launched from small water craft manually (i.e., Mk 18 Mod 1 Swordfish UUV)
Lightweight vehicle class: up to 500 lb displacement, 20–40 hours endurance; launched from RHIB using launch/retriever system or by cranes from surface ships (i.e., Mk 18 Mod 2 Kingfish UUV)
Heavyweight vehicle class: up to 3000 lb displacement, 40–80 hours endurance, launched from submarines
Large vehicle class: up to 10 long tons displacement; launched from surface ships and submarines
Vehicle designs
Hundreds of different AUVs have been designed over the past 50 or so years, but only a few companies sell vehicles in any significant numbers. There are around 10 companies that sell AUVs on the international market, including Kongsberg Maritime, Hydroid (now a wholly owned subsidiary of Kongsberg Maritime), Bluefin Robotics, Teledyne Gavia (previously known as Hafmynd), International Submarine Engineering (ISE) Ltd, Atlas Elektronik, and OceanScan.
Vehicles range in size from man portable lightweight AUVs to large diameter vehicles of over 10 metres length. Large vehicles have advantages in terms of endurance and sensor payload capacity; smaller vehicles benefit significantly from lower logistics (for example: support vessel footprint; launch and recovery systems).
Some manufacturers have benefited from domestic government sponsorship including Bluefin and Kongsberg. The market is effectively split into three areas: scientific (including universities and research agencies), commercial offshore (oil and gas, etc.) and military application (mine countermeasures, battle space preparation). The majority of these roles utilize a similar design and operate in a cruise (torpedo-type) mode. They collect data while following a preplanned route at speeds between 1 and 4 knots.
Commercially available AUVs include various designs, such as the small REMUS 100 AUV originally developed by Woods Hole Oceanographic Institution in the US and now produced commercially by Hydroid, Inc. (a wholly owned subsidiary of Kongsberg Maritime); the larger HUGIN 1000 and 3000 AUVs developed by Kongsberg Maritime and Norwegian Defence Research Establishment; the Bluefin Robotics 12-and-21-inch-diameter (300 and 530 mm) vehicles and the International Submarine Engineering Ltd. Most AUVs follow the traditional torpedo shape as this is seen as the best compromise between size, usable volume, hydrodynamic efficiency and ease of handling. There are some vehicles that make use of a modular design, enabling components to be changed easily by the operators.
The market is evolving and designs are now following commercial requirements rather than being purely developmental. Upcoming designs include hover-capable AUVs for inspection and light-intervention (primarily for the offshore energy applications), and hybrid AUV/ROV designs that switch between roles as part of their mission profile. Again, the market will be driven by financial requirements and the aim to save money and expensive ship time.
Today, while most AUVs are capable of unsupervised missions, most operators remain within range of acoustic telemetry systems in order to maintain a close watch on their investment. This is not always possible. For example, Canada has recently taken delivery of two AUVs (ISE Explorers) to survey the sea floor underneath the Arctic ice in support of their claim under Article 76 of the United Nations Convention of the Law of the Sea. Also, ultra-low-power, long-range variants such as underwater gliders are becoming capable of operating unattended for weeks or months in littoral and open ocean areas, periodically relaying data by satellite to shore, before returning to be picked up.
As of 2008, a new class of AUVs are being developed, which mimic designs found in nature. Although most are currently in their experimental stages, these biomimetic (or bionic) vehicles are able to achieve higher degrees of efficiency in propulsion and maneuverability by copying successful designs in nature. Two such vehicles are Festo's AquaJelly (AUV) and the EvoLogics BOSS Manta Ray.
Sensors
AUVs carry sensors to navigate autonomously and map features of the ocean. Typical sensors include compasses, depth sensors, sidescan and other sonars, magnetometers, thermistors and conductivity probes. Some AUVs are outfitted with biological sensors including fluorometers (also known as Chlorophyll sensors), turbidity sensors, and sensors to measure pH, and amounts of dissolved oxygen.
A demonstration at Monterey Bay in California in September 2006 showed that a 21-inch (530 mm) diameter AUV can tow a 400 feet (120 m) long hydrophone array while maintaining a 6-knot (11 km/h) cruising speed.
Navigation
Radio waves cannot penetrate water very far, so as soon as an AUV dives it loses its GPS signal. Therefore, a standard way for AUVs to navigate underwater is through dead reckoning. Navigation can however be improved by using an underwater acoustic positioning system. When operating within a net of sea floor deployed baseline transponders this is known as LBL navigation. When a surface reference such as a support ship is available, ultra-short baseline (USBL) or short-baseline (SBL) positioning is used to calculate where the sub-sea vehicle is relative to the known (GPS) position of the surface craft by means of acoustic range and bearing measurements. To improve estimation of its position, and reduce errors in dead reckoning (which grow over time), the AUV can also surface and take its own GPS fix. Between position fixes and for precise maneuvering, an Inertial Navigation System on board the AUV calculates through dead reckoning the AUV position, acceleration, and velocity. Estimates can be made using data from a Inertial Measurement Unit, and can be improved by adding a Doppler Velocity Log (DVL), which measures the rate of travel over the sea/lake floor. Typically, a pressure sensor measures the vertical position (vehicle depth), although depth and altitude can also be obtained from DVL measurements. These observations are filtered to determine a final navigation solution.
Propulsion
There are a couple of propulsion techniques for AUVs. Some of them use a brushed or brush-less electric motor, gearbox, Lip seal, and a propeller which may be surrounded by a nozzle or not. All of these parts embedded in the AUV construction are involved in propulsion. Other vehicles use a thruster unit to maintain the modularity. Depending on the need, the thruster may be equipped with a nozzle for propeller collision protection or to reduce noise submission, or it may be equipped with a direct drive thruster to keep the efficiency at the highest level and the noises at the lowest level. Advanced AUV thrusters have a redundant shaft sealing system to guarantee a proper seal of the robot even if one of the seals fails during the mission.
Underwater gliders do not directly propel themselves. By changing their buoyancy and trim, they repeatedly sink and ascend; airfoil "wings" convert this up-and-down motion to forward motion. The change of buoyancy is typically done through the use of a pump that can take in or push out water. The vehicle's pitch can be controlled by changing the center of mass of the vehicle. For Slocum gliders this is done internally by moving the batteries, which are mounted on a screw. Because of their low speed and low-power electronics, the energy required to cycle trim states is far less than for regular AUVs, and gliders can have endurances of months and transoceanic ranges.
Power
Most AUVs in use today are powered by rechargeable batteries (lithium ion, lithium polymer, nickel metal hydride etc.), and are implemented with some form of Battery Management System. Some vehicles use primary batteries which provide perhaps twice the endurance—at a substantial extra cost per mission. A few of the larger vehicles are powered by aluminum based semi-fuel cells, but these require substantial maintenance, require expensive refills and produce waste product that must be handled safely. An emerging trend is to combine different battery and power systems with supercapacitors.
Source from Wikipedia
History
The first AUV was developed at the Applied Physics Laboratory at the University of Washington as early as 1957 by Stan Murphy, Bob Francois and later on, Terry Ewart. The "Special Purpose Underwater Research Vehicle", or SPURV, was used to study diffusion, acoustic transmission, and submarine wakes.
Other early AUVs were developed at the Massachusetts Institute of Technology in the 1970s. One of these is on display in the Hart Nautical Gallery in MIT. At the same time, AUVs were also developed in the Soviet Union (although this was not commonly known until much later).
Applications
Until relatively recently, AUVs have been used for a limited number of tasks dictated by the technology available. With the development of more advanced processing capabilities and high yield power supplies, AUVs are now being used for more and more tasks with roles and missions constantly evolving.
Commercial
The oil and gas industry uses AUVs to make detailed maps of the seafloor before they start building subsea infrastructure; pipelines and sub sea completions can be installed in the most cost effective manner with minimum disruption to the environment. The AUV allows survey companies to conduct precise surveys of areas where traditional bathymetric surveys would be less effective or too costly. Also, post-lay pipe surveys are now possible, which includes pipeline inspection. The use of AUVs for pipeline inspection and inspection of underwater man-made structures is becoming more common.
Research
A University of South Florida researcher deploys Tavros02, a solar-powered "tweeting" AUV (SAUV)
Scientists use AUVs to study lakes, the ocean, and the ocean floor. A variety of sensors can be affixed to AUVs to measure the concentration of various elements or compounds, the absorption or reflection of light, and the presence of microscopic life. Examples include conductivity-temperature-depth sensors (CTDs), fluorometers, and pH sensors. Additionally, AUVs can be configured as tow-vehicles to deliver customized sensor packages to specific locations.
Hobby
Many roboticists construct AUVs as a hobby. Several competitions exist which allow these homemade AUVs to compete against each other while accomplishing objectives. Like their commercial brethren, these AUVs can be fitted with cameras, lights, or sonar. As a consequence of limited resources and inexperience, hobbyist AUVs can rarely compete with commercial models on operational depth, durability, or sophistication. Finally, these hobby AUVs are usually not oceangoing, being operated most of the time in pools or lake beds. A simple AUV can be constructed from a microcontroller, PVC pressure housing, automatic door lock actuator, syringes, and a DPDT relay. Some participants in competitions create open-source designs.
Illegal drug traffic
Submarines that travel autonomously to a destination by means of GPS navigation have been made by illegal drug traffickers.
Air crash investigations
Autonomous underwater vehicles, for example AUV ABYSS, have been used to find wreckages of missing airplanes, e.g. Air France Flight 447, and the Bluefin-21 AUV was used in the search for Malaysia Airlines Flight 370.
Military applications
The U.S. Navy Unmanned Undersea Vehicle (UUV) Master Plan identified the following UUV's missions:
Intelligence, surveillance, and reconnaissance
Mine countermeasures
Anti-submarine warfare
Inspection/identification
Oceanography
Communication/navigation network nodes
Payload delivery
Information operations
Time-critical strike
The Navy Master Plan divided all UUVs into four classes:
Man-portable vehicle class: 25–100 lb displacement; 10–20 hours endurance; launched from small water craft manually (i.e., Mk 18 Mod 1 Swordfish UUV)
Lightweight vehicle class: up to 500 lb displacement, 20–40 hours endurance; launched from RHIB using launch/retriever system or by cranes from surface ships (i.e., Mk 18 Mod 2 Kingfish UUV)
Heavyweight vehicle class: up to 3000 lb displacement, 40–80 hours endurance, launched from submarines
Large vehicle class: up to 10 long tons displacement; launched from surface ships and submarines
Vehicle designs
Hundreds of different AUVs have been designed over the past 50 or so years, but only a few companies sell vehicles in any significant numbers. There are around 10 companies that sell AUVs on the international market, including Kongsberg Maritime, Hydroid (now a wholly owned subsidiary of Kongsberg Maritime), Bluefin Robotics, Teledyne Gavia (previously known as Hafmynd), International Submarine Engineering (ISE) Ltd, Atlas Elektronik, and OceanScan.
Vehicles range in size from man portable lightweight AUVs to large diameter vehicles of over 10 metres length. Large vehicles have advantages in terms of endurance and sensor payload capacity; smaller vehicles benefit significantly from lower logistics (for example: support vessel footprint; launch and recovery systems).
Some manufacturers have benefited from domestic government sponsorship including Bluefin and Kongsberg. The market is effectively split into three areas: scientific (including universities and research agencies), commercial offshore (oil and gas, etc.) and military application (mine countermeasures, battle space preparation). The majority of these roles utilize a similar design and operate in a cruise (torpedo-type) mode. They collect data while following a preplanned route at speeds between 1 and 4 knots.
Commercially available AUVs include various designs, such as the small REMUS 100 AUV originally developed by Woods Hole Oceanographic Institution in the US and now produced commercially by Hydroid, Inc. (a wholly owned subsidiary of Kongsberg Maritime); the larger HUGIN 1000 and 3000 AUVs developed by Kongsberg Maritime and Norwegian Defence Research Establishment; the Bluefin Robotics 12-and-21-inch-diameter (300 and 530 mm) vehicles and the International Submarine Engineering Ltd. Most AUVs follow the traditional torpedo shape as this is seen as the best compromise between size, usable volume, hydrodynamic efficiency and ease of handling. There are some vehicles that make use of a modular design, enabling components to be changed easily by the operators.
The market is evolving and designs are now following commercial requirements rather than being purely developmental. Upcoming designs include hover-capable AUVs for inspection and light-intervention (primarily for the offshore energy applications), and hybrid AUV/ROV designs that switch between roles as part of their mission profile. Again, the market will be driven by financial requirements and the aim to save money and expensive ship time.
Today, while most AUVs are capable of unsupervised missions, most operators remain within range of acoustic telemetry systems in order to maintain a close watch on their investment. This is not always possible. For example, Canada has recently taken delivery of two AUVs (ISE Explorers) to survey the sea floor underneath the Arctic ice in support of their claim under Article 76 of the United Nations Convention of the Law of the Sea. Also, ultra-low-power, long-range variants such as underwater gliders are becoming capable of operating unattended for weeks or months in littoral and open ocean areas, periodically relaying data by satellite to shore, before returning to be picked up.
As of 2008, a new class of AUVs are being developed, which mimic designs found in nature. Although most are currently in their experimental stages, these biomimetic (or bionic) vehicles are able to achieve higher degrees of efficiency in propulsion and maneuverability by copying successful designs in nature. Two such vehicles are Festo's AquaJelly (AUV) and the EvoLogics BOSS Manta Ray.
Sensors
AUVs carry sensors to navigate autonomously and map features of the ocean. Typical sensors include compasses, depth sensors, sidescan and other sonars, magnetometers, thermistors and conductivity probes. Some AUVs are outfitted with biological sensors including fluorometers (also known as Chlorophyll sensors), turbidity sensors, and sensors to measure pH, and amounts of dissolved oxygen.
A demonstration at Monterey Bay in California in September 2006 showed that a 21-inch (530 mm) diameter AUV can tow a 400 feet (120 m) long hydrophone array while maintaining a 6-knot (11 km/h) cruising speed.
Navigation
Radio waves cannot penetrate water very far, so as soon as an AUV dives it loses its GPS signal. Therefore, a standard way for AUVs to navigate underwater is through dead reckoning. Navigation can however be improved by using an underwater acoustic positioning system. When operating within a net of sea floor deployed baseline transponders this is known as LBL navigation. When a surface reference such as a support ship is available, ultra-short baseline (USBL) or short-baseline (SBL) positioning is used to calculate where the sub-sea vehicle is relative to the known (GPS) position of the surface craft by means of acoustic range and bearing measurements. To improve estimation of its position, and reduce errors in dead reckoning (which grow over time), the AUV can also surface and take its own GPS fix. Between position fixes and for precise maneuvering, an Inertial Navigation System on board the AUV calculates through dead reckoning the AUV position, acceleration, and velocity. Estimates can be made using data from a Inertial Measurement Unit, and can be improved by adding a Doppler Velocity Log (DVL), which measures the rate of travel over the sea/lake floor. Typically, a pressure sensor measures the vertical position (vehicle depth), although depth and altitude can also be obtained from DVL measurements. These observations are filtered to determine a final navigation solution.
Propulsion
There are a couple of propulsion techniques for AUVs. Some of them use a brushed or brush-less electric motor, gearbox, Lip seal, and a propeller which may be surrounded by a nozzle or not. All of these parts embedded in the AUV construction are involved in propulsion. Other vehicles use a thruster unit to maintain the modularity. Depending on the need, the thruster may be equipped with a nozzle for propeller collision protection or to reduce noise submission, or it may be equipped with a direct drive thruster to keep the efficiency at the highest level and the noises at the lowest level. Advanced AUV thrusters have a redundant shaft sealing system to guarantee a proper seal of the robot even if one of the seals fails during the mission.
Underwater gliders do not directly propel themselves. By changing their buoyancy and trim, they repeatedly sink and ascend; airfoil "wings" convert this up-and-down motion to forward motion. The change of buoyancy is typically done through the use of a pump that can take in or push out water. The vehicle's pitch can be controlled by changing the center of mass of the vehicle. For Slocum gliders this is done internally by moving the batteries, which are mounted on a screw. Because of their low speed and low-power electronics, the energy required to cycle trim states is far less than for regular AUVs, and gliders can have endurances of months and transoceanic ranges.
Power
Most AUVs in use today are powered by rechargeable batteries (lithium ion, lithium polymer, nickel metal hydride etc.), and are implemented with some form of Battery Management System. Some vehicles use primary batteries which provide perhaps twice the endurance—at a substantial extra cost per mission. A few of the larger vehicles are powered by aluminum based semi-fuel cells, but these require substantial maintenance, require expensive refills and produce waste product that must be handled safely. An emerging trend is to combine different battery and power systems with supercapacitors.
Source from Wikipedia
Swarm robotics
Swarm robotics is an approach to the coordination of multiple robots as a system which consist of large numbers of mostly simple physical robots. It is supposed that a desired collective behavior emerges from the interactions between the robots and interactions of robots with the environment. This approach emerged on the field of artificial swarm intelligence, as well as the biological studies of insects, ants and other fields in nature, where swarm behaviour occurs.
Definition
The research of swarm robotics is to study the design of robots, their physical body and their controlling behaviours. It is inspired but not limited by the emergent behaviour observed in social insects, called swarm intelligence. Relatively simple individual rules can produce a large set of complex swarm behaviours. A key-component is the communication between the members of the group that build a system of constant feedback. The swarm behaviour involves constant change of individuals in cooperation with others, as well as the behaviour of the whole group.
Unlike distributed robotic systems in general, swarm robotics emphasizes a large number of robots, and promotes scalability, for instance by using only local communication. That local communication for example can be achieved by wireless transmission systems, like radio frequency or infrared.
Goals and applications
Miniaturization and cost are key factors in swarm robotics. These are the constraints in building large groups of robots; therefore the simplicity of the individual team member should be emphasized. This should motivate a swarm-intelligent approach to achieve meaningful behavior at swarm-level, instead of the individual level.
Much research has been directed at this goal of simplicity at the individual robot level. Being able to use actual hardware in research of Swarm Robotics rather than simulations allows researchers to encounter and resolve many more issues and broaden the scope of Swarm Research. Thus, development of simple robots for Swarm intelligence research is a very important aspect of the field. The goals include keeping the cost of individual robots low to allow scalability, making each member of the swarm less demanding of resources and more power/energy efficient.
One such swarm system is the LIBOT Robotic System that involves a low cost robot built for outdoor swarm robotics. The robots are also made with provisions for indoor use via Wi-Fi, since the GPS sensors provide poor communication inside buildings. Another such attempt is the micro robot (Colias), built in the Computer Intelligence Lab at the University of Lincoln, UK. This micro robot is built on a 4 cm circular chassis and is low-cost and open platform for use in a variety of Swarm Robotics applications.
Advantages and disadvantages
The most frequently cited benefits are:
low cost for more extensive coverage;
a redundancy capacity (if one of the robots fails due to a failure, blockage, etc. another robot can take steps to troubleshoot or replace it in its task).
the ability to cover a large area. Duarte & al. have for example shown (via a simulation applied to the case of the island of Lampedusa) in 2014 that a swarm of 1000 small aquatic drones dispersed at sea from bases could in 24 hours make a surveillance report on a maritime band 20 km long;
To date, swarms of robots can only perform relatively simple tasks, they are often limited by their need for energy. More generally, the difficulties of interoperability when one wants to associate robots of nature and of different origins are also very limiting.
Properties
Unlike most distributed robotic systems, swarm robotics insists on a large number of robots 6 and promotes scaling, for example the use of local communications in the form of infrared or wireless.
These systems are expected to have at least the following three properties:
robustness, which implies the ability of the swarm to continue to function despite the failures of certain individuals and / or changes that may occur in the environment;
flexibility, which implies a capacity to propose solutions adapted to the tasks to be performed;
the "scaling", which implies that the swarm must function regardless of its size (from a certain minimum size).
According to Sahin (2005) and Dorigo (2013) in a swarm robotic system, in the swarm:
Each robot is autonomous;
robots are usually able to locate themselves relative to their closest neighbors (relative positioning) and sometimes in the global environment, even if some systems try to do without this data;
the robots can act (eg to modify the environment, to cooperate with another robot);
The detection and communication capabilities of robots between them are local (lateral) and limited;
the robots are not connected to a centralized control; they do not have the global knowledge of the system in which they cooperate;
the robots cooperate to perform a given task;
emerging phenomena global behaviors can thus appear.
Applications
Potential applications for swarm robotics are many. They include tasks that demand miniaturization (nanorobotics, microbotics), like distributed sensing tasks in micromachinery or the human body. One of the most promising uses of swarm robotics is in disaster rescue missions. Swarms of robots of different sizes could be sent to places rescue workers can't reach safely, to detect the presence of life via infra-red sensors. On the other hand, swarm robotics can be suited to tasks that demand cheap designs, for instance mining or agricultural foraging tasks.
More controversially, swarms of military robots can form an autonomous army. U.S. Naval forces have tested a swarm of autonomous boats that can steer and take offensive actions by themselves. The boats are unmanned and can be fitted with any kind of kit to deter and destroy enemy vessels.
Most efforts have focused on relatively small groups of machines. However, a swarm consisting of 1,024 individual robots was demonstrated by Harvard in 2014, the largest to date.
Another large set of applications may be solved using swarms of micro air vehicles, which are also broadly investigated nowadays. In comparison with the pioneering studies of swarms of flying robots using precise motion capture systems in laboratory conditions, current systems such as Shooting Star can control teams of hundreds of micro aerial vehicles in outdoor environment using GNSS systems (such as GPS) or even stabilize them using onboard localization systems where GPS is unavailable. Swarms of micro aerial vehicles have been already tested in tasks of autonomous surveillance, plume tracking, and reconnaissance in a compact phalanx. Numerous works on cooperative swarms of unmanned ground and aerial vehicles have been conducted with target applications of cooperative environment monitoring, convoy protection, and moving target localization and tracking.
Drone displays
A drone display commonly uses multiple, lighted drones at night for an artistic display.
In Popular Culture
A major subplot of Disney's Big Hero involved the use of swarms of microbots to form structures.
Research
They cover many topics including:
software and software enhancement;
improving the robots themselves. In 2010, two Swiss researchers from Lausanne (Floreano & Keller) proposed to draw inspiration from Darwinian (adaptive) selection to develop robots;
the ability to evolve in 3 dimensions (in the air for a fleet of aerial drones, or underwater for a swarm of underwater robots), for example for the study of the dynamics of water bodies and marine currents;
improving their ability to cooperate with each other or with other types of robots;
on swarm behavior assessment (video tracking is essential to study swarm behavior in a systematic way, even if other methods exist, such as the recent development of ultrasonic tracking.) Further research is needed to establish a methodology suitable for the design and reliable prediction of swarms when only the traits of individuals are known);
comparing the respective advantages and disadvantages of top-down and bottom-up approaches.
Source from Wikipedia
Definition
The research of swarm robotics is to study the design of robots, their physical body and their controlling behaviours. It is inspired but not limited by the emergent behaviour observed in social insects, called swarm intelligence. Relatively simple individual rules can produce a large set of complex swarm behaviours. A key-component is the communication between the members of the group that build a system of constant feedback. The swarm behaviour involves constant change of individuals in cooperation with others, as well as the behaviour of the whole group.
Unlike distributed robotic systems in general, swarm robotics emphasizes a large number of robots, and promotes scalability, for instance by using only local communication. That local communication for example can be achieved by wireless transmission systems, like radio frequency or infrared.
Goals and applications
Miniaturization and cost are key factors in swarm robotics. These are the constraints in building large groups of robots; therefore the simplicity of the individual team member should be emphasized. This should motivate a swarm-intelligent approach to achieve meaningful behavior at swarm-level, instead of the individual level.
Much research has been directed at this goal of simplicity at the individual robot level. Being able to use actual hardware in research of Swarm Robotics rather than simulations allows researchers to encounter and resolve many more issues and broaden the scope of Swarm Research. Thus, development of simple robots for Swarm intelligence research is a very important aspect of the field. The goals include keeping the cost of individual robots low to allow scalability, making each member of the swarm less demanding of resources and more power/energy efficient.
One such swarm system is the LIBOT Robotic System that involves a low cost robot built for outdoor swarm robotics. The robots are also made with provisions for indoor use via Wi-Fi, since the GPS sensors provide poor communication inside buildings. Another such attempt is the micro robot (Colias), built in the Computer Intelligence Lab at the University of Lincoln, UK. This micro robot is built on a 4 cm circular chassis and is low-cost and open platform for use in a variety of Swarm Robotics applications.
Advantages and disadvantages
The most frequently cited benefits are:
low cost for more extensive coverage;
a redundancy capacity (if one of the robots fails due to a failure, blockage, etc. another robot can take steps to troubleshoot or replace it in its task).
the ability to cover a large area. Duarte & al. have for example shown (via a simulation applied to the case of the island of Lampedusa) in 2014 that a swarm of 1000 small aquatic drones dispersed at sea from bases could in 24 hours make a surveillance report on a maritime band 20 km long;
To date, swarms of robots can only perform relatively simple tasks, they are often limited by their need for energy. More generally, the difficulties of interoperability when one wants to associate robots of nature and of different origins are also very limiting.
Properties
Unlike most distributed robotic systems, swarm robotics insists on a large number of robots 6 and promotes scaling, for example the use of local communications in the form of infrared or wireless.
These systems are expected to have at least the following three properties:
robustness, which implies the ability of the swarm to continue to function despite the failures of certain individuals and / or changes that may occur in the environment;
flexibility, which implies a capacity to propose solutions adapted to the tasks to be performed;
the "scaling", which implies that the swarm must function regardless of its size (from a certain minimum size).
According to Sahin (2005) and Dorigo (2013) in a swarm robotic system, in the swarm:
Each robot is autonomous;
robots are usually able to locate themselves relative to their closest neighbors (relative positioning) and sometimes in the global environment, even if some systems try to do without this data;
the robots can act (eg to modify the environment, to cooperate with another robot);
The detection and communication capabilities of robots between them are local (lateral) and limited;
the robots are not connected to a centralized control; they do not have the global knowledge of the system in which they cooperate;
the robots cooperate to perform a given task;
emerging phenomena global behaviors can thus appear.
Applications
Potential applications for swarm robotics are many. They include tasks that demand miniaturization (nanorobotics, microbotics), like distributed sensing tasks in micromachinery or the human body. One of the most promising uses of swarm robotics is in disaster rescue missions. Swarms of robots of different sizes could be sent to places rescue workers can't reach safely, to detect the presence of life via infra-red sensors. On the other hand, swarm robotics can be suited to tasks that demand cheap designs, for instance mining or agricultural foraging tasks.
More controversially, swarms of military robots can form an autonomous army. U.S. Naval forces have tested a swarm of autonomous boats that can steer and take offensive actions by themselves. The boats are unmanned and can be fitted with any kind of kit to deter and destroy enemy vessels.
Most efforts have focused on relatively small groups of machines. However, a swarm consisting of 1,024 individual robots was demonstrated by Harvard in 2014, the largest to date.
Another large set of applications may be solved using swarms of micro air vehicles, which are also broadly investigated nowadays. In comparison with the pioneering studies of swarms of flying robots using precise motion capture systems in laboratory conditions, current systems such as Shooting Star can control teams of hundreds of micro aerial vehicles in outdoor environment using GNSS systems (such as GPS) or even stabilize them using onboard localization systems where GPS is unavailable. Swarms of micro aerial vehicles have been already tested in tasks of autonomous surveillance, plume tracking, and reconnaissance in a compact phalanx. Numerous works on cooperative swarms of unmanned ground and aerial vehicles have been conducted with target applications of cooperative environment monitoring, convoy protection, and moving target localization and tracking.
Drone displays
A drone display commonly uses multiple, lighted drones at night for an artistic display.
In Popular Culture
A major subplot of Disney's Big Hero involved the use of swarms of microbots to form structures.
Research
They cover many topics including:
software and software enhancement;
improving the robots themselves. In 2010, two Swiss researchers from Lausanne (Floreano & Keller) proposed to draw inspiration from Darwinian (adaptive) selection to develop robots;
the ability to evolve in 3 dimensions (in the air for a fleet of aerial drones, or underwater for a swarm of underwater robots), for example for the study of the dynamics of water bodies and marine currents;
improving their ability to cooperate with each other or with other types of robots;
on swarm behavior assessment (video tracking is essential to study swarm behavior in a systematic way, even if other methods exist, such as the recent development of ultrasonic tracking.) Further research is needed to establish a methodology suitable for the design and reliable prediction of swarms when only the traits of individuals are known);
comparing the respective advantages and disadvantages of top-down and bottom-up approaches.
Source from Wikipedia
订阅：
博文 (Atom)
-
Proletarian Mother David Alfaro Siqueiros 1931 From the collection of Museo Nacional de Arte David Alfaro Siqueiros was an outstanding soci...
-
Traditional games activities is a popular form of entertainment as well as an essential part of Bangladeshi culture. Icim bicim Itching i...
-
Objective abstraction was a British art movement. Between 1933 and 1936 several artists later associated with the Euston Road School produce...
