The illusion that reactive and hierarchical planning methods are at odds with each other needs to be dropped. By exploiting each method's strengths, a synthesis of hierarchical and reactive paradigms can yield robust, flexible, and generalizable navigation. Psychological and neuroscientific studies support this claim. 1. Introduction The integration of knowledge-based navigational path planning and reactive navigation requires the confrontation of many difficult problems. It can be seen that each of these methods addresses different subsets of the complexities inherent in intelligent navigation. It is our contention that neither navigational approach is entirely satisfactory when taken in isolation, but rather that both must be taken into account for the production of an intelligent, robust, and flexible system. Navigational path planning without consideration for the difficult issues of plan execution leads to restricted usage in very narrow problem domains and/or extremely brittle m...

Our experience over the years with different architectures and planning systems for mobile robots has led us to a distributed approach where an arbiter receives votes for and against commands from each subsystem and decides upon the course of action which best satisfies the current goals and constraints of the system. Centralized arbitration of votes from distributed, independent decision-making processes provides coherent, rational, goal-directed behavior while preserving real-time responsiveness to its immediate physical environment. The Distributed Architecture for Mobile Navigation (DAMN) has been successfully used to integrate various independently developed subsystems, providing systems that perform road following, cross-country navigation, or teleoperation while avoiding obstacles and meeting mission objectives. Examples of implemented systems are given. Further research will seek to more rigorously define the behavior of the system. Keywords: mobile robots, architecture, behav...

We describe a reactive robotic control system which incorporates aspects of machine learning to improve the system's ability to successfully navigate in unfamiliar environments. This system overcomes limitations of completely reactive systems by exercising on-line performance enhancement without the need for high level planning. The results of extensive simulation studies using the learning enhanced reactive controller are presented. 1. Introduction Reactive robotic control systems [1,5,11] have produced significant results in generating intelligent robotic action when compared to previous efforts. These systems typically decompose actions into behaviors in order to produce rapid real-time sensory response. How these systems can adapt ongoing behaviors to the environment is an important first step in addressing learning in reactive control. The particular approach described in this paper enables a reactive control system to adapt its behavior based on recent experience. When there are...

This article presents a new line of research investigating on-line learning mechanisms for autonomous intelligent agents. We discuss a case-based method for dynamic selection and modification of behavior assemblages for a navigational system. The casebased reasoning module is designed as an addition to a traditional reactive control system, and provides more flexible performance in novel environments without extensive high-level reasoning that would otherwise slow the system down. The method is implemented in the ACBARR (A Case-BAsed Reactive Robotic) system, and evaluated through empirical simulation of the system on several different environments, including "box canyon" environments known to be problematic for reactive control systems in general.

This paper describes a series of experiments that were performed on the Rocky III robot. 1 Rocky III is a small autonomous rover capable of navigating through rough outdoor terrain to a predesignated area, searching that area for soft soil, acquiring a soil sample, and depositing the sample in a container at its home base. The robot is programmed according to a reactive behaviorcontrol paradigm using the ALFA programming language. This style of programming produces robust autonomous performance while requiring significantly less computational resources than more traditional mobile robot control systems. The code for Rocky III runs on an 8-bit processor and uses about 10k of memory. Introduction The research described in this paper is motivated by NASA's planetary rover program. A planetary rover would be used on missions to deploy instruments and collect samples outside of the immediate area surrounding a lander. As science instruments get smaller and more sensitive, the size and s...

This paper presents a self-improving reactive control system for autonomous robotic navigation. The navigation module uses a schema-based reactive control system to perform the navigation task. The learning module combines case-based reasoning and reinforcement learning to continuously tune the navigation system through experience. The case-based reasoning component perceives and characterizes the system's environment, retrieves an appropriate case, and uses the recommendations of the case to tune the parameters of the reactive control system. The reinforcement learning component refines the content of the cases based on the current experience. Together, the learning components perform on-line adaptation, resulting in improved performance as the reactive control system tunes itself to the environment, as well as on-line case learning, resulting in an improved library of cases that capture environmental regularities necessary to perform on-line adaptation. The system is extensively evaluated through simulation studies using several performance metrics and system configurations. 1 Introduction

To increase speed and reliability of operation, multiple computers are replacing uniprocessors and wired-logic controllers in modern robots and industrial control systems. However, performance increases are not attained by such hardware alone. The operating software controlling the robots or control systems must exploit the possible parallelism of various control tasks in order to perform the necessary computations within given real-time and reliability constraints. Such software consists of both control programs written by application programmers and operating system software offering means of task scheduling, intertask communication, and device control. The Generalized Executive for real-time Multiprocessor applications (GEM) is an operating system that addresses several requirements of operating software. First, when using GEM, programmers can select one of two different types of tasks differing in size, called processes and microprocesses. Second, the scheduling calls offered by GEM permit the implementation of several models of task interaction. Third, GEM supports multiple models of communication with a parameterized communication mechanism. Fourth, GEM is closely coupled to prototype real-time programming environments that provide programming support for the models of computation offered by the operating system. GEM is being used on a multiprocessor with robotics application software of substantial size and complexity.

Most current architectures for autonomous robots are based on a decomposition of the control problem into small units of control, or behaviors. While this decomposition has a number of advantages, it brings about the problem of having to coordinate the execution of different units in order to obtain a globally coherent behavior. In this paper, we discuss how fuzzy logic can be used, and has been used, to address this problem. 1. Introduction Procs. of the 6th IEEE Int. Conf. on Fuzzy Systems (Barcelona, SP, july 1997) 573-578 The goal of autonomous robotics is to build physical systems that accomplish useful tasks without human intervention in real-world, unmodified environments --- that is, in environments that have not been specifically engineered for the robot. A major challenge of autonomous robotics is the large amounts of uncertainty that characterizes real-world environments. On the one hand, we cannot have exact and complete prior knowledge of these environments: many detail...

This paper presents a self-improving reactive control system for autonomous robotic navigation. The navigation module uses a schemabased reactive control system to perform the navigation task. The learning module combines case-based reasoning and reinforcement learning to continuously tune the navigation system through experience. The case-based reasoning component perceives and characterizes the system’s environment, retrieves an appropriate case, and uses the recommendations of the case to tune the parameters of the reactive control system. The reinforcement learning component refines the content of the cases based on the current experience. Together, the learning components perform on-line adaptation, resulting in improved performance as the reactive control system tunes itself to the environment, as well as on-line learning, resulting in an improved library of cases that capture environmental regularities necessary to perform on-line adaptation. The system is extensively evaluated through simulation studies using several performance metrics and system configurations.

A control system architecture is described for an autonomous flying vehicle. The vehicle, equipped with fourteen sensors, uses a model helicopter as an airframe. The control system utilizes these sensors to a) remain aloft and in stable flight, b) navigate to a target and c) manipulate a physical object. The overall approach to the problem is based on a behavioral paradigm. The key contribution of this paper is the demonstration of a situated agent under these severe circumstances; as the craft is airborne, it is in constant risk of crashing. Unlike terrestrial mobile robots, the craft must constantly make sound decisions to maintain its integrity. 1.