Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. of the 11th National ConferenceonArtificial Intelligence, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....notion of executable abstract plans regardless of their level of detail is made possible by using plans as suggestions that direct the base level execution mechanism but do not impel a particular behavior. This idea was promoted by Agre and Chapman [1990] and was experimentally supported by Gat [1992]. Uncertainty alone makes it impossible to use plans except as a guidance mechanism. Performance with perfect vision What is the performance of the abstract planner First, I will examine the performance under the assumption of perfect domain description. Then, I will examine the effect of ....

E. Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the Tenth National Conference on Artificial Intelligence, pp. 809--815, San Jose, California, 1992.

.... many three tiered archi tectures fit the first case described in the introduc tion and have a degenerated deliberative layer but a powerful sequencing layer, for example, one based on RAPS [Firby87] The planners of some of these robot architectures run asynchronously with the control loop [Gat91 ], whereas the plmners of others run synchronously with the control loop [Bon97] Similarly, the plmners of some of these robot architectures run continuously [Sten95] Lyons95] whereas the planners of others run only from time to time [Bon97] The planner of our robot architecture runs ....

E. Gat, "Integrating planning and reacting in a heterogeneous asynchronous architecture for mobile robots", SIGART Bulletin, Vol. 2, 1991, pp. 70-74.

.... since an iterative goal refinement process can be stopped at various points along the way and provide useful (but incomplete) advice to 2A similar approach has been proposed by Agre and Chapman s plans as communications theory [1] It has been experimentally verified in the ATLANTIS architecture [5]. the execution architecture. Obviously, the advice provided after short deliberation may be less detailed and less helpful than a complete plan, but it may be the right thing to do under real time pressure. 4 History based evaluation We propose to analyze and evaluate alternative planning ....

E. Gat. Integrating Planning and Reacting in a Hetero- geneous Asynchronous Architecture for Controlling RealWorld Mobile Robots. Proceedings of the Tenth National Conference on Artificial Intelligence, pp. 809-815, San Jose, California, 1992.

....simple failure recovery procedures, particularly for ac tions that have common and known failures. For example, an action might simply be repeated, as in the navigateToGoal command. These procedures resemble schemas [Georgeff Ingrand, 1989, Hotmann, Meier, Schloen, 1991] or RAPs [Firby, 1989, Gat, 1992, Pellet al... 1997] in that they specify how to execute the action, what to monitor in the environment, and some recovery procedures. ROGUE s procedures, however, do not contain complex recovery or monitoring procedures, such as when they have different costs or probabilities, since we feel that ....

Gat, E. (1992). Integrating planning and re- acting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the Tenth National Conference on Artificial Intelligence (AAAI-92). pp. 809-815.

....the route from sensor data to actuator settings; the condition action pairs are like reflexes in biological systems. Reactive robots have good real time characteristics, which is an important advantage over classically controlled robots because An oncoming truck waits for no theorem prover [[6]] They can keep pace with changes in a dynamic environment, since the path from change in sensor reading to change in actuator setpoint is short, providing a direct coupling of perception to action. 14 2.3.2 Example architecture The Niche Robot Architecture, Miller [ 5] This is a very ....

....moved towards removal of the idea of success and failure , and a change of focus from process steps to time extended activities. This architecture again retains the need to operate at a symbolic level, with all the difficulties that entails (see Section 2.4.7, page 22) 2.5. 6 ATLANTIS, Gat [[6]] This is a combination of a reactive control substrate with a classical planning system. Planning is seen as an attempt to transform one world state to another using operators which map on to associated physical actions when executed. In the classical approach, operators are executed in an ....

Gat, E.: Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Controlling Real-World Mobile Robots, Proceedings of the 10th National Conference on Artificial Intelligence, pp. 809-815, MIT Press, July 1992.

....them. Experiments indicate the feasibility of the scheme and the sophistication possible even with this minimalist approach. 1. INTRODUCTION Mobile robots in a number of research laboratories have demonstrated excellent capabilities, from performing self protecting behaviors to navigation [2,3,5,14,16,18]. While useful tasks may be executed without internal representations [6] flexible, goal oriented navigation requires an internal representation (map) of the environment [4,14,16] A map can assume different forms according to the relationships it establishes among locations, ranging from ....

.... laboratories have demonstrated excellent capabilities, from performing self protecting behaviors to navigation [2,3,5,14,16,18] While useful tasks may be executed without internal representations [6] flexible, goal oriented navigation requires an internal representation (map) of the environment [4,14,16]. A map can assume different forms according to the relationships it establishes among locations, ranging from detailed geometry to abstract graphs [11] Autonomous execution of complex tasks involves external sensing devices. Robot capabilities hinge upon adequate sensing, however, more sensing ....

E. Gat, "Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Controlling Real-World Mobile Robots", AAAI-92.

....away important execution details that are needed to transition between the plan states. All of these di#culties have driven robotics researchers to produce their own plan 4 representations and system architectures. One popular approach to plan representation is that of making reactive plans [29, 38] in which the appropriate course of action to take to accomplish a goal depends on the robot s situational context. Such plans allow robots to be robust in the face of many sensor and actuator failures and also to adapt to (or take advantage of) unforeseen changes in the environment. Typically, ....

Erann Gat, Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots, Tenth National Conference on Artificial Intelligence, 1992.

....distinct schools. Architectures in the situated behaviour school [Brooks, 1986; Agre and Chapman, 1987; Kaelbling and Rosenschein, 1990] typically allow frequent changes in the actions of the robot, yet restrict the allowable computational models. The planning school [Nilsson, 1984; Firby, 1992; Gat, 1992] allows unrestricted computational models, yet the commitment to arbitrary length plans hinders the ability DataCube RGB Camera (Single CCD) Available Transputers Vision Engine Transputer Network 1 2 n User Nodes: Reasoning Control Radio Transmitter Soccer Field User UNIX ....

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In AAAI-92, pages 809--815, 1992.

....planner. Thus, we may expect non deterministic planners to be able to handle domains with larger state spaces than RL. An alternative to RL and non deterministic planning is to interleave planning with execution. This approach has been widely used in non deterministic robotic domains e.g. [33, 32, 60, 35]. A group of planners suitable for this purpose is action selectors based on heuristic search [45, 5] Several bdd based planning approaches has been developed for non deterministic domains. They can be divided into two groups according to the form of the synthesized plan. The first group keeps ....

E. Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the 10th National Conference on Artificial Intelligence (AAAI'92), pages 809--815. AAAI Press, 1992.

....soccer domain the two joint actions achieving the goal shown in Figure 17 were extracted from the universal plan in less than 0.001 seconds. 7 Previous Work Recurrent approaches performing planning in parallel with execution have been widely used in non deterministic robotic domains (e.g. [21, 19, 44, 25]) A group of planners suitable for recurrent planning is action selectors based on heuristic search [30, 4] The min max lrta algorithm [30] can generate suboptimal plans in non deterministic domains through a search and execution iteration. The search is based on a heuristic goal distance ....

E. Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the 10'th National Conference on Artificial Intelligence (AAAI'92), pages 809--815. MIT Press, 1992.

....where the situation changes dynamically. ROBBIE addresses this by combining deliberative and reactive planning, and by adding new features of situations through introspective learning as they turn out to be important. Other work has addressed the integration of deliberative and reactive planning: Gat #1992# implements a framework based on Firby s RAP system in ATLANTIS, in which the reasoning process can request advice from higher level deliberative planning processes. Nourbakhsh, Powers, Birch#eld #1995# describe an alternative approach that maintains a set of the possible locations in ....

....the time that must be spent deliberating about the details of the plan, while at the same time ensuring that the reactive planner has enough guidance to avoid problems due to its lack of long term reasoning. Combining deliberative planning with reactive planning has been studied by various people #Gat, 1992; Nourbakhsh et al. 1995#, but ROBBIE is the #rst instance we know of a case based planner being combined with a reactive planner. Future directions: While ROBBIE s planning components perform well together, further work could be done to balance the power of the reactive planner against that of ....

[Article contains additional citation context not shown here]

Gat, E. #1992#. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-worl mobile robots. In Proceedings, Tenth National Conference on Arti#cial Intelligence,pp.809#815.

....behavior modes by subsuming the more general goals of lower level behaviors. In binary decision networks, the system changes behavior as a result of direct sensory input. In all cases, the system achieves complex goals by switching control modes to achieve different subtasks at different times [3, 11, 10]. 2 RAP Executor Active Tracking Routines Primary Action Routines RAP Library World Model World Tasks Requests Requests Results Figure 1: A Goal Directed, Modular Control System ffl Control system modes can be targeted using attributes of the world that can be tracked in real ....

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Tenth National Conference on Artificial Intelligence, San Jose, CA, July 1992. AAAI.

....system s flow of control. To Appear in the Second International Conference on AI Planning Systems, June 1994. 1 Introduction Recently, AI researchers have proposed several different mechanisms for programming robots reactively. These include collections of behaviors [2] schemas [1] routines [9], and reflexes [15] Many details differ between these proposals, particularly in the area of philosophical commitment, but they share the common idea that the actual behavior of the robot at any given moment is the result of a set of interacting processes acting on input from the environment. ....

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Tenth National Conference on Artificial Intelligence, San Jose, CA, July 1992. AAAI.

No context found.

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. of the 11th National ConferenceonArtificial Intelligence, 1992.

No context found.

Gat, E. (1992). Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the AAAI Conference.

No context found.

Erann Gat, Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real world mobile robots, Procs. of AAAI-92.

No context found.

Gat, E., "Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Controlling Real-World Mobile Robots," Proc. of AAAI-92, pp. 809-815, Cambridge, MA: MIT Press, 1992.

No context found.

E. Gat, "Integrating planning and reacting in a heterogeneous asynchronous architecture for mobile robots," SIGART Bulletin, Vol. 2, 1991, pp. 70-74.

No context found.

E. Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proc. of AAAI-91, San Jose, CA, 1992.

No context found.

E. Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the Tenth National Conference on Artificial Intelligence (AAAI-92), pages 802--815, 1992.

No context found.

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. In Proceedings of the Tenth National Conference on Artificial Intelligence (AAAI'92), 1992.

No context found.

E. Gat. Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Controlling Real-World Mobile Robots. In Proceedings of AAAI-92, pp. 809--815, 1992.

No context found.

Erann Gat. Integrating planning and reacting in a heterogeneous asynchronous architecture for controlling real-world mobile robots. Proceedings of the 8th National Conference on Artificial Intelligence, pages 809--815, 1992.

No context found.

Gat, E. #1992# `Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Controlling Real-World Mobile Robots.' Proceedings of the Tenth National Conference on Arti#cial Intelligence #AAAI-92#, 809#815.

No context found.

E. Gat, Integrating Planning and Reacting in a Heterogeneous Asynchronous Architecture for Mobil Robots, SIGART Bulletin 2(1991) 70-74.

First 50 documents  Next 50

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