This paper describes an implementation of the 3T robot architecture which has been under development for the last eightyears. The architecture uses three levels of abstraction and description languages whichare compatible between levels. The makeup of the architecture helps to coordinate planful activities with real-time behaviors for dealing with dynamic environments. In recent years, other architectures have been created with similar attributes but two features distinguish the 3T architecture: 1) a variety of useful software tools have been created to help implement this architecture on multiple real robots;, and 2) this architecture, or parts of it, have been implemented on a varietyofvery different robot systems using different processors, operating systems, effectors and sensor suites.

Renewed motives for space exploration have inspired NASA to work toward the goal of establishing a virtual presence in space, through heterogeneous effets of robotic explorers. Information technology, and Artificial Intelligence in particular, will play a central role in this endeavor by endowing these explorers with a form of computational intelligence that we call remote agents. In this paper we describe the Remote Agent, a specific autonomous agent architecture based on the principles of model-based programming, on-board deduction and search, and goal-directed closed-loop commanding, that takes a significant step toward enabling this future. This architecture addresses the unique characteristics of the spacecraft domain that require highly reliable autonomous operations over long periods of time with tight deadlines, resource constraints, and concurrent activity among tightly coupled subsystems. The Remote Agent integrates constraint-based temporal planning and scheduling, robust multi-threaded execution, and model-based mode identification and reconfiguration. The demonstration of the integrated system as an on-board controller for Deep Space One, NASA's rst New Millennium mission, is scheduled for a period of a week in late 1998. The development of the Remote Agent also provided the opportunity to reassess some of AI's conventional wisdom about the challenges of implementing embedded systems, tractable reasoning, and knowledge representation. We discuss these issues, and our often contrary experiences, throughout the paper.

This paper describes an extension to the rap system task-net semantics and representation language to enable the effective control of continuous processes. The representation addresses the problems of synchronizing plan expansion with events in the world, coping with multiple, non-deterministic task outcomes, and the description of a simple form of clean-up task. It is also pointed out that success and failure need no special place in a task network representation. Success and failure are really messages about the execution system's knowledge and do not explicitly define that 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 philosop...

AI planning agents are goal-directed: success is measured in terms of whether or not an input goal is satisfied, and the agent's computational processes are driven by those goals. A decision-theoretic agent, on the other hand, has no explicit goals--- success is measured in terms of its preferences or a utility function that respects those preferences. The two approaches have complementary strengths and weaknesses. Symbolic planning provides a computational theory of plan generation, but under unrealistic assumptions: perfect information about and control over the world and a restrictive model of actions and goals. Decision theory provides a normative model of choice under uncertainty, but offers no guidance as to how the planning options are to be generated. This paper unifies the two approaches to planning by describing utility models that support rational decision making while retaining the goal information needed to support plan generation. We develop an extended model of goals tha...

This paper describes the New Millennium Remote Agent #NMRA# architecture for autonomous spacecraft control systems. This architecture integrates traditional real-time monitoring and control with constraintbased planning and scheduling, robust multi-threaded execution, and model-based diagnosis and recon#guration.

The Hap reactive agent architecture provides many of the same mechanisms to authors as other goal-directed reactive architectures: arbitrary demons, multiple active goals and situated, runtime interpretation of plans. It also provides three unique features: convenient mechanisms for taking advantage of opportunities as they arise and for changing the agent's course of action when unfortunate events dictate; explicit mechanisms for allowing easy extension of the plan library; and flexible mechanisms for determining the success of goals without necessarily requiring the author to explicitly characterize the criteria for success. Hap was designed as part of the Oz project which is concerned with developing a rich framework for both interactive fiction and virtual realities. Oz is intended to provide a human user with the experience of living in a dramatically interesting simulated world which includes simulated intelligent and emotional agents.

Vision systems that have successfully supported nontrivial tasks have invariably taken advantage of constraints derived from the task and environment to increase reliability and lower the complexity of perception. We propose that it is possible to build a general purpose vision system, that is, one that can support a wide variety of tasks, and take advantage of such constraints. The central idea within our proposed architecture is the reactive skill. Skills are concurrent control routines assembled at run time using instructions from a symbolic execution system. Visual modules are used as resources in the construction of these skills. Skills control the agent as continuous feedback loops but are constructed using discrete, symbolic instructions. The key to general-purpose vision is the ability to parametrize the primitive elements of the vision system and to compose visual and control routines in a variety of ways. We demonstrate the architecture in the context of an implemented examp...

In this paper we address the problem of controlling the execution of a query optimizer. We describe a control for the optimization process that is based on planning. The controller described here is a goal-directed planner that intermingles planning with the execution of query transformations, and uses execution results to direct further planning of optimizer processing. We describe this control in the context of the Epoq extensible architecture. Epoq is an approach to extensible query optimization that integrates specialized rewrite strategies through its extensible control mechanism. This paper describes our planning-based approach to extensible control and illustrates it with a simple example. 1 Introduction Optimization of a query is inherently a process of searching the space of expressions equivalent to the query. Typically, a given optimizer can only visit some portion of this space, since the set of transformation rules is usually incomplete, the cost of optimization must be ...

The methodological underpinnings of AI are slowly changing. Benchmarks, testbeds, and controlled experimentation are becoming more common. While we are optimistic that this change can solidify the science of AI, we also recognize a set of difficult issues concerning the appropriate use of this methodology. We discuss these issues as they relate to research on agent design. We survey existing testbeds for agents, and argue for appropriate caution in their use. We end with a debate on the proper role of experimental methodology in the design and validation of planning agents. Hanks was supported in part by NSF grants IRI-9008670 and IRI-9206733. y Pollack was supported by the Air Force Office of Scientific Research, Contracts F49620-91-C-0005 and F49620-92-J-0422, by the Rome Laboratory (RL) and the Defense Advanced Research Projects Agency, Contract F30602-93-C-0038, and by an NSF Young Investigator's Award, IRI-9258392 z Cohen was supported in part by the Defense Advanced Researc...

This paper is about the interface between continuous and symbolic robot control. We advocate describing continuous actions and their related sensing strategies as situation specific activities, which can be manipulated by a symbolic reactive planner. The approach addresses the issues involved in turning symbolic actions into continuous activities, and using task specific sensing routines to support those activities. Situation specific activities help preserve the convenient fiction of "primitive actions" for use in planning without requiring that they all be programmed into the control system in advance. We demonstrate the utility of this architecture with an object tracking example. A control system is presented that can be reconfigured by a the rap reactive executor to achieve different tasks. We show how this system allows us to build interchangeable tracking activities that use different sensing /action feedback loops in different situations. First Internation...