Brooks, R. A. and Connell, J. H. (1986). Asynchronous distributed control system for a mobile robot. In Proceedings of the SPIE's Cambridge Symposyum on Optical and Optoelectronic Engineering, pages 77--84.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....nature of actual ultrasonic range information makes it difficult to reproduce test runs or compare them with each other. The VFF algorithm, however, works i T i 1 ( T) i 1 8 (7) equally well with real obstacles, as is demonstrated in the last experimental result in this paper (Fig. 8) All of the following examples and plots have been obtained from actual runs of our robot (with its sensors disabled) Although the maximum speed of V =0.78 m sec was used, the max average speed in each run was lower, since the algorithm reduces the robot s speed under certain conditions (as ....

....forces by the directional cosine (cos ) of the two vectors, F and V, and r using the product as follows: F = wF (1 w) F ( cos ) 8) r r r where F is the adjusted sum of the repulsive r forces and w is a weighting factor that was set to w = 0.25 in our system. The directional cosine in Eq. 8 is computed by where V , V = x and y components of velocity vector x y V F , F = x and y components of the sum of the rx ry repulsive forces, F r V V max for F r 0 (i.e. in the absence of obstacles) V max (1 cos ) for F r 0 10 (10) The effect of this damping method is ....

[Article contains additional citation context not shown here]

Brooks, R. A. and Connell, J. H., "Asynchronous Distributed Control System for a Mobile Robot", Proceedings of the SPIE, Vol. 727, Mobile Robots, 1986, pp. 77-84.

....represent obstacles and the robot s path is generated off line. We use the term probability in the literal sense of likelyhood. 1 Page 5 While each of the above methods features valuable refinements, none have been implemented on a mobile robot with real sensory data. By contrast, Brooks [8,9] and Arkin [1] use a potential field method on experimental mobile robots (equipped with a ring of ultrasonic sensors) Brooks implementation treats each ultrasonic range reading as a repulsive force vector. If the magnitude of the sum of the repulsive forces exceeds a certain threshold, the ....

....grid) and can influence the path of the vehicle immediately. A single range reading will have only minor influence on the path, while repeated range readings in a confined area (cluster) will cause a more drastic change of direction for the vehicle. The force field method developed by Brooks [8,9] and the similar method developed by Arkin [1] do function in experimentalreal time systems, using actual sensory data [8,9] However, these methods are somewhat oversimplified, since a threshold determines if an object is at a safe distance or too close. In the latter case, and because of the ....

[Article contains additional citation context not shown here]

Brooks, R. A. and Connell, J. H., "Asynchronous Distributed Control System for a Mobile Robot." Proceedings of the SPIE, Mobile Robots, Vol. 727, 1987, pp. 77-84.

....the issue of achieving global objectives through local rules (Lewis Bekey 1992, Agah Bekey 1994b) Behavior based systems combine the best properties of the two extremes of planning and reactive systems without resorting to hybridizing control. They are an extension of reactive architectures (Brooks 1986a) and embody some of the properties of reactive systems, but their computation is not limited to look up and they are able to adapt over time because behaviors can be employed to store various forms of state and representations (Matari c 1997) Our previous work has used them in navigation, ....

....interface to the designer and improve possibility of spin offs and ease of technology transfer. Our approach combines basis behaviors into powerful dynamic networks behavior networks which supersede the fixed layered approach of previous behavioral robotics designs such as sub sumption (Brooks 1986b) This model can combine low level reactive behaviors ( avoid obstacle ) with goal directed behaviors ( follow waypoints to target ) and adaptive behaviors ( learn features of hazardous obstacles ) without any centralized control. Each basis behavior of a standard library will be implemented as ....

[Article contains additional citation context not shown here]

Brooks, R. A. & Connell, J. H. (1986), Asynchronous distributed control system for a mobile robot, in `SPIE's Cambridge Symposium on Optical and Optoelecronic Engineering', Cambridge, MA, pp. 77--84.

....efforts of many institutions, due to their immediate applicability in a wide variety of tasks such as space missions, operations in hazardous environments, and military operations. Leading research programs on mobile robots and related topics include (1) The MIT Mobile Robot Lab [3] 4] 5] 6] [7] [8] 9] 22] 24] 25] 26] 35] 2) The Shakey Robot of SRI [31] 32] 33] 3 A preliminary version of this paper was presented at the 5th IEEE International Symposium on Intelligent Control, September 1990, Philadelphia, Pennsylvania. This work was supported in part by the National Science ....

R. A. Brooks and J. H. Connel, "Asynchronous distributed control system for a mobile robot," in SPIE Cambridge Symp. Optical and Opto-Electronic Eng. Proc., Vol. 727, 1986, pp. 77-84.

....strategies that combine two or more of these approaches. For an excellent, detailed discussion, see Mataric [101] 3.3. 1 Purely Reactive Strategies The simplest control strategy, conceptually, is the purely reactive one implemented as a set of condition action pairs (e.g. Brooks and Connell [34, 41] and Agre and Chapman [3] The biggest advantage of these systems is their speed, due to their minimal computation. They consist of a set of purely reactive rules, containing little or no internal state. They simply use sensor readings to index into this set, selecting and executing the ....

Rodney A. Brooks and Jonathan H. Connell. Asynchronous distributed control system for a mobile robot. In SPIE's Cambridge Symposium on Optical and Opto-Electronic Engineering Proceedings, volume 727, pages 77--84, October 1986.

....environment and act using stimulus response type of behavior: they respond to the present state of the environment in which they are embedded. Thus, reactive agents follow simple patterns of behavior which can easily be programmed. Studies on reactive agents can be traced to the works of [3] [4], 5] and [6] The figure 1 below illustrates the MUTANT agents model. stimulus action Environment sensors actuators Behavioral Model a mutant Figure 1. the MUTANT agents model 3.1. The behavioral model In the agent domain, a behavior based system generally relies on coordination of some ....

R. Brooks and J.H. Connell, Asynchronous Distributed Control System for a Mobile Robot, SPIE 727, 1986.

....a fast heuristic for the purpose of reconfiguring the system in response to failed valves, modal processes solve a more restricted problem for the purpose of propagating mode changes imperatively. ACTs can also be prioritized, allowing behavioral composition similar to the subsumption architecture [2]. While handlers are allowed to change only those modes that are local to their process, ACTs allow the local effects to be propagated to other processes globally, as well as customizing the behavior of individual processes. SomecommonlyusedACTsareshowninTable1. The most common way control is ....

R. A. Brooks and J. H. Connell. Asynchronous distributed control system for a mobile robot. In Proceedings of the SPIE - The International Society for Optical Engineering, volume 727, pages 77--84, 1987.

.... is shared by many researchers in AI and psychology (Brooks [Bro91] Meas [Mae89] Agre and Chapman [AC87] Hewitt [Hew91] Minsky [Min86] Beer [Bee90] Braitenberg [Bra84] Brooks and his colleagues did very interesting work on building artificial creatures [BCN88, Bro88, Con90] Brooks [Bro86, BC86] proposed a robust, layered control system for mobile robots, called the subsumption architecture. Unlike the traditional decomposition of a mobile robot control system into functional modules, Brooks decomposed a mobile robot control system into task achieving behaviors. Maes [Mae89] suggested ....

R. A. Brooks and J. H. Connell. Asynchronous distributed control system for a mobile robot. SPIE Mobile Robots, 727, 1986.

....to achieve reactivity outside of the classical paradigm. In 1987, Agre and Chapman [1] described a system which achieved a modest level of performance at playing a highly dynamic video game, while using a control regimen that was completely driven by responses to environmental influences. Brooks [5, 8] began a disciplined approach to the design and construction of mobile robots that dogmatically rejected the techniques of classical planning. Sanborn and Hendler [36] advocated a system where several cooperating processes, making strictly local computations and working without a strong model of ....

R. A. Brooks and J. H. Connell, "Asynchronous Distributed Control System for a Mobile Robot," in SPIE Vol. 727 Mobile Robots, pp. 77--84, 1986.

....world may or may not be physical. Various attempts at achieving real time performance have been proposed. Perhaps the most prominent are purely reactive bottom up approaches which implement the agent s control strategy as a collection of preprogrammed condition action pairs with minimal state (Brooks Connell 1986, Agre Chapman 1987, Connell 1990) These systems maintain no internal models and perform no search, but simply look up and command the appropriate action for each set of sensor readings. They rely on a direct coupling between sensing and action, and fast feedback from the environment. Purely ....

....cases, the low level reactive process takes care of the immediate safety of the robot, while the higher level uses the planner to select action sequences. Behavior based approaches are an extension of reactive systems that also fall between the purely reactive and the planner based extremes (Brooks 1986, Maes 1989) Although often confused in the literature, behavior based strategies are strictly more powerful than purely reactive approaches since they have no fundamental limitations on internal state. While behavior based systems embody some of the properties of reactive systems, and usually ....

[Article contains additional citation context not shown here]

Brooks, R. A. & Connell, J. H. (1986), Asynchronous Distributed Control System for a Mobile Robot, in `SPIE', Cambridge, Massachusetts.

....the expected sensory information and on the relative importance of that information to the overall system operation at that point in time. The expression of the above fusion ideas using = s should be straightforward. This is not the case with methodologies like Brooks sumbsumption architecture, [7], where (as it will be explained in the next section) even a dynamic priority scheme cannot be supported. 3.3 Behavioral Specification using the = model Building autonomous creatures has been a particularly interesting and challenging area in robotics research. This goal has led to research in ....

....behaviors [11] that can deal with multiple goals and multiple sensory information. In this section, we show how the = model (and language) can be used to specify such behaviors in a natural, concise and elegant way. We contrast the = framework with Brooks subsumption architecture [7]. In particular, we show that the = model is more general and more expressive; it subsumes the subsumption architecture. In [7] Brooks proposes the subsumption architecture as a methodology for specifying and building complex control systems. This architecture suggests the use of a vertical ....

[Article contains additional citation context not shown here]

Brooks R., and Connell, J., "Asynchronous Distributed Control System for a Mobile Robot", SPIE Proceedings, Vol. 727, October 1986.

....tasks which output differing, and usually conflicting, motion requests. These motion request must be handled by an operating system of some kind to produce a unique motion command. We describe a simple, extensible, operating system, similar in flavor to Brooks subsumption architecture [2], which can be embedded on the same chip as the sensors and the sensor processing circuitry. 2 Integrating Sensors and Sensor Processing Circuitry The first stage in any sensori motor system is processing the raw sensor data so that the information relevant to the motor task is extracted from the ....

Brooks R., and Connell, J., "Asynchronous Distributed Control System for a Mobile Robot", SPIE Proceedings, Vol. 727, October 1986.

....for real time response. The basic control design approaches can be broadly divided into four types. These are defined by [Mataric 92a] and we briefly reiterate them and their shortcomings here. The purely reactive approaches use a mapping from sensor sets to associated actions; a set of rules [Brooks 87] The planner based strategies originated with the symbolic AI community and employ a sense plan act cycle. The plan stage uses cognitive techniques to reason about a symbolic world model. There also exist hybrid systems which employ reactive components beneath planner based systems to provide ....

Brooks, Rodney A. and Connell, Jonathan H., "Asynchronous Distributed Control System for a Mobile Robot", SPIE's Cambridge Symposium on Optical and Opto-Electronic Engineering Proceedings, Vol. 727, October, 1986, 77-84.

....to perform a concrete task. The basic control design approaches can be broadly divided into four types. These are defined by [Mataric 92a] and we briefly reiterate them and their shortcomings here. The purely reactive approaches use a mapping from sensor sets to associated actions; a set of rules [Brooks 87] The planner based strategies originated with the symbolic AI community and employ a sense plan act cycle. The plan stage uses cognitive techniques to reason about a symbolic world model. There also exist hybrid systems which employ reactive components beneath planner based systems to provide ....

....action plans, which is currently being implemented. One major problem to be solved in robotics is, given a robot with a repertoire of basic behaviours which behaviour should be selected next Various action selection mechanisms have been proposed, such as Rodney Brooks subsumption architecture [Brooks 87] We have developed a mechanism based loosely on Pattie Maes spreading activation scheme [Maes 90a] extended to add integrated learning and adapted to a behaviour based framework. This is beyond the scope of this paper, please refer to [Jung 97b] Briefly, in simplified form, behaviours are ....

Brooks, Rodney A. and Connell, Jonathan H., "Asynchronous Distributed Control System for a Mobile Robot", SPIE's Cambridge Symposium on Optical and Opto-Electronic Engineering Proceedings, Vol. 727, October, 1986, 77-84.

....Also, low level behaviors need not be concerned with high level information but can focus on tasks such as move forward 1 meter . The information needed at each level is therefore limited, making the search space smaller. Brooks subsumption architecture is a hierarchical behavior based approach [Brooks and Connell, 1986], Brooks, 1985] Lower levels control instinctive behaviors and higher levels control tasks that are regarded as more abstract. In a mobile robot for example, a low level behavior might be obstacle avoidance, while a high level behavior controls path planning. Each behavior consists of a fixed ....

Brooks, Rodney A. and Connell, Jonathan H. 1986. Asynchronous distributed control system for a mobile robot. SPIE 727:77--84.

....and evaluated empirically through extensive simulation. 1 A notable exception is recent work on reinforcement learning (see discussion below) In contrast, the research presented in this paper is based on work in case based reasoning. 2 Overview Reactive robotic control systems [Arkin, 1989; Brooks, 1986; Kaelbling, 1986; Payton, 1986] have produced impressive results in the area of generating intelligent robotic action. Unlike traditional approaches to robot control, these systems typically decompose actions into simple behaviors in order to produce rapid real time response to the environment. ....

....systems would perform poorly. 3 Background and related research Before presenting the technical details of our system, we discuss previous research in reactive control and in machine learning that forms the basis for our work. 3. 1 Perception and reactive control Reactive control [Arkin, 1989; Brooks, 1986; Brooks, 1989; Kaelbling, 1986; Payton, 1986] is concerned with how to coordinate multiple motor behaviors. It is characterized by a tight coupling between perception and action with little or no intervening representation. This results in systems which do not perform detailed planning but are ....

[Article contains additional citation context not shown here]

R. Brooks and J. Connell. Asynchronous Distributed Control System for a Mobile Robot. In W. Wolfe and N. Marquina, editors, Proceedings of the SPIE, Volume 727: Mobile Robots, pages 77--84, Bellingham, WA, 1986.

....decision making and reasoning about its surroundings. To handle the challenges of real time sensory input and processing in an unknown and constantly changing environment, reactive architectures have been used. One reactive approach is the subsumption behavior architecture developed by Brooks [2, 3]. Independent, decision making processes, called behaviors, execute concurrently. Each behavior receives sensor data relevant to its decision making needs and produces a control output based on its desired action. Independent behaviors may present conflicting control actions that must be resolved. ....

R. A. Brooks and J. H. Connell. Asynchronous distributed control system for mobile robot control. In Proceedings of SPIE, volume 727, pages 77--84, 1986.

....robot goes about its other tasks. As the number of objects the robot might need to locate grows, this examination would become computationally intractable. Remembering these locations in any sort of Cartesian coordinate system is also often impractical, not only because of sensor and effector error[Brooks, 1987], but also because objects commonly are moved. The ability to perform object search might help avoid these difficulties. Rather than attempt to remember specific object locations, a robot could rely on its object search system to find objects only as they are needed. 2 Efficiency is crucial for ....

....of goal directed or task oriented perception, a popular approach to the design of sensory systems in which the idea is to preprogram as much knowledge about where to look or what to look for. For example, Connell constructed a robot that roams an area searching for and collecting soda cans[Brooks and Connell, 1987; Connell, 1989] Because soda cans are usually on tables, the robot was designed only to look for cans at a fixed height. This research concentrated more on the issues involved in constructing robots with a subsumption architecture than on object search issues. Its treatment of issues such as ....

Rodney A. Brooks and Jonathan H. Connell, "Asynchronous distributed control system for a mobile robot," In Mobile Robots: Proc. SPIE 727, 1987.

....on the Input Output Timed Automaton (IOTA read yota ) model. We have used the IOTA framework to model robotics applications [Bestavros:90b] The tasks involved in a robotic application are diverse (vision [BBN:86] motion control [Brockett:88] high level planning and behavioral specification [Brooks:86], etc. and usually make use of very different resources (special purpose image processors, tailor made controllers and drivers, general purpose processors, etc. Furthermore, these tasks interact in a non trivial way. Being able to specify and verify such complex systems in a single ....

Rodney Brooks and Jonathan Connell, "Asynchronous Distributed Control System for a Mobile Robot", SPIE Proceedings, Vol. 727, October 1986.

....believed to be a better solution. As an example of this approach consider the field of robotics, where a bottom up design philosophy has initiated research in behavioristic systems where small systems are created which gradually are expanded into more complex entities [Brooks, 1985, Brooks, 1991, Brooks and Connell, 1987, Connell, 1989] These systems are then analyzed and it is hoped that this in time may evolve into a mathematical (or CHAPTER 1. INTRODUCTION 5 just a set of known rules) definition of how to design intelligent systems. The mathematical formulation of system design has, however, still to be ....

....elaborate attention selection strategy than what is embedded in most current camera head vision systems, that typically rely on more or less advanced event detectors. There exist numerous examples on how selective processing may speed up and make the visual search a tractable problem. For example [Brooks and Connell, 1987, Connell, 1989] have build robots based on the subsumption architecture. In a demonstrator, they showed that this architecture was applicable for building a robot that could find and remove soda cans from tables. To limit the search space strict constraints were enforced on where soda cans ....

Rodney A. Brooks and Jonathan H. Connell. Asynchronous distributed control system for a mobile robot. In Mobile Robots: Proceedings of SPIE 727. SPIE, 1987.

.... the work in reactive robotics and in particular on the Subsumption Architecture [12] which achieves rapid real time responses by embedding the robot s controller into a collection of preprogrammed parallel condition action rules, or reflexes, with minimal internal state (e.g. if bumped, stop ) [15, 3]. In contrast to these socalled bottom up systems, traditional AI deliberative planner based systems are top down, and require the robot to perform a serial sequence of processing sense plan act steps (e.g. combine the sensory data into a model of the world, then use the planner to find a path ....

Rodney A. Brooks and Jonathan H. Connell. Asynchronous distributed control system for a mobile robot. In SPIE's Cambridge Symposium on Optical and Optoelecronic Engineering, pages 77--84, Cambridge, MA, 26--31 October 1986.

....approaches for achieving real time performance in autonomous agents have been proposed. Purely reactive bottom up approaches are featured in various implemented systems. They embed the agent s control strategy into a collection of preprogrammed condition action pairs with minimal internal state (Brooks Connell 1986, Agre Chapman 1987, Connell 1990) Reactive systems maintain no internal models and perform no search. Typically, they apply a simple functional mapping between stimuli and appropriate responses, usually in the form of a lookup, whether it be in a table, a set of reactive rules, a simple ....

....reactive process takes care of the immediate safety of the agent, while the higher level uses the planner to select action sequences. 3 Behavior Based Approaches Behavior based approaches are an extension of reactive architectures and also fall between purely reactive and planner based extremes (Brooks 1986, Maes 1989) Although often conflated in the literature, reactive and behavior based systems are fundamentally different. While behavior based systems embody some of the properties of reactive systems, and usually contain reactive components, their computation is not limited to look up and ....

[Article contains additional citation context not shown here]

Brooks, R. A. & Connell, J. H. (1986), Asynchronous Distributed Control System for a Mobile Robot, in `SPIE', Cambridge, MA.

....up all the balls from the playing field. In this approach, the robot has to constantly poll its sensors to find the next goal state (the next ball to pick up) until it has picked up all the balls. The subsumption architecture was the first attempt at a generalized formalization of this approach [11]. In the subsumption architecture, each robot is associated with a repertoire of behaviors that it can exhibit. Goal recognition is performed via polling the environment and selecting the next goal (and the corresponding behavior that achieves this goal) Any change in the environment can cause ....

.... this goal) Any change in the environment can cause the currently executing behavior to be subsumed by a new behavior (e.g. a deer wandering through the forest may be searching for food but seeing a hunter will cause it to hide in a cave) The subsumed behavior may be resumed at a later time [11]. 1.2.3 Path Planning Once the goals of a robot have been identified, the next task is to navigate towards them. The two major motion strategies used for mobile robot navigation are reactive and planning [27] In reactive navigation, only the goal is known (and sometimes its location) However, ....

[Article contains additional citation context not shown here]

R.A. Brooks and J.H. Connell. Asynchronous Distributed Control System for a Mobile Robot. Proceedings of SPIE, vol. 727, Oct 1986, pp 77-84.

....predefined schemas as necessary to produce more complex behaviors. Such apparent complexity arises both from the ability of schemas to use other schemas and from the parallel actions of independent schemas. Another significant reactive approach is the subsumption architecture developed by Brooks [9, 10, 11]. Multiple levels of competence are defined, connecting input and output in a layered system. Higher levels of competence inhibit or subsume all lower levels, and the hardware usually provides direct support for this subsumption characteristic by implementing each level within its own processing ....

Brooks, R. A., and Connell, J. H. "Asynchronous Distributed Control System for a Mobile Robot," Proc. SPIE, vol. 727, 1986, pp. 77-84.

....avoidance layer. The physical robot did not therefore remain true to the desires of the upper layer. The upper layer had to watch what happened in the world, through odometry, in order to understand what was really happening in the lower control layers, and send down correction signals. In [7] we described an alternate set of layers for the robot Allen. 3.2 Tom and Jerry Tom and Jerry [8] were two identical robots built to demonstrate just how little raw computation is necessary to support the subsumption architecture. A three layer subsumption programme was implemented, yet al..l data ....

R.A. Brooks and J.H. Connell, "Asynchronous Distributed Control System for a Mobile Robot", SPIE Vol. 727, Mobile Robots, Cambridge, MA, 77-84, November 1986.

....avoidance layer. The physical robot did not therefore remain true to the desires of the upper layer. The upper layer had to watch what happened in the world, through odometry, in order to understand what was really happening in the lower control layers, and send down correction signals. In [9] we described an alternate set of layers for the robot Allen. 4.2. Tom and Jerry Tom and Jerry [14] were two identical robots built to demonstrate just how little raw computation is necessary to support the subsumption architecture. A three layer subsumption program was implemented, yet al..l data ....

Rodney A. Brooks and Jonathan H. Connell, Asynchronous distributed control system for a mobile robot, SPIE Vol. 727 Mobile Robots, Cambridge, MA (November 1986) 77-84.

....without representation Rodney A. Brooks MIT Artificial Intelligence Laboratory, 545 Technology Square, Rm. 836, Cambridge, MA 02139, USA Received September 1987 Brooks, R.A. Intelligence without representation, Artificial Intelligence 47 (1991) 139 159. This report describes research done at the Artificial Intelligence Laboratory of the Massachusetts Institute of ....

....without representation Rodney A. Brooks MIT Artificial Intelligence Laboratory, 545 Technology Square, Rm. 836, Cambridge, MA 02139, USA Received September 1987 Brooks, R.A. Intelligence without representation, Artificial Intelligence 47 (1991) 139 159. This report describes research done at the Artificial Intelligence Laboratory of the Massachusetts Institute of ....

[Article contains additional citation context not shown here]

R.A. Brooks and J.H. Connell, Asynchronous distributed control system for a mobile robot, in: Proceedings SPIE, Cambridge, MA (1986) 77-84.

....that all human behavior is simply the external expression of a seething mass of rather independent behaviors without any central control or representations of the world. Maybe there is only chaos from which order appears to emerge. THE SUBSUMPTION ARCHITECTURE The subsumption architecture Brooks (1986) is a parallel and distributed computation formalism for connecting sensors to actuators in robots. A traditional way of describing these connections would be to say the subsumption architecture provides a way of writing intelligent control programs for mobile robots. One writes a subsumption ....

....Finally we outline a subsumption program for a complex visually guided creature (named Seymour) that is currently under development. Allen Our first robot, Allen, had sonar distance sensors and odometry onboard and used an offboard lisp machine to simulate the subsumption architecture. In Brooks (1986) we described three layers of control implemented in the subsumption architecture. The wiring diagram is shown in Figure 8.3. The first layer let the robot avoid both static and dynamic obstacles; Allen would happily sit in the middle of a morn until approached, then scurry away, avoiding ....

[Article contains additional citation context not shown here]

Brooks, R. A., & Connell, 1. H. (1986). Asynchronous Distributed Control System for a Mobile Robot. SPIE Vol. 727 Mobile Robots, 77-84.

No context found.

Brooks, R. A. and Connell, J. H. (1986). Asynchronous distributed control system for a mobile robot. In Proceedings of the SPIE's Cambridge Symposyum on Optical and Optoelectronic Engineering, pages 77--84.

No context found.

Brooks, R. A. and Connell, J. H. (1986). Asynchronous distributed control system for a mobile robot. In Proceedings of the SPIE's Cambridge Symposyum on Optical and Optoelectronic Engineering, pages 77--84.

No context found.

R. A. Brooks and J. H. Connell. Asynchronous distributed control system for a mobile robot. In SPIE, Cambridge, MA, USA, 1986.

No context found.

Rodney A. Brooks and Jonathan H. Connell. Asynchronous distributed control system for a mobile robot. In SPIE's CambridgeSymposium on Optical and Opto-Electronic Engineering Proceedings, volume 727, pages 77--84, October 1986.

No context found.

Rodney A. Brooks and Jonathan H. Connell. Asynchronous distributed control system for a mobile robot. In Cambridge Symposium on Optical and Optoelectronic Engineering. SPIE, October 1986.

No context found.

R.A.Brooks and J.H.Connell. Asynchronous distributed control system for a mobile robot. In Mobile Robots:Proc. SPIE 727, California, 1987.

No context found.

R. A. Brooks and J. H. Connell, Asynchronous distributed control system for a mobile robot, in Mobile Robots: Proc. SPIE 727, California, 1987.

No context found.

Brooks R. and Connel J., Asynchronous Distributed Control System for a Mobile Robot, Proc. Eds. W. Wolfe and N. Marquina SPIE Mobile Robots, Bellingham, Wash (1986), pp. 77-84.

No context found.

R.A. Brooks and J.H. Connell. "Asynchronous Distributed Control System for a Mobile Robot," Proceedings of SPIE, vol. 727, Oct 1986, pp 77-84.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC