Accession Number : ADA294103
Title : Using a Genetic Algorithm to Learn Behaviors for Autonomous Vehicles,
Corporate Author : NAVY CENTER FOR APPLIED RESEARCH IN ARTIFICIAL INTELLIGENCE WASHINGTON DC
Personal Author(s) : Schultz, Alan C. ; Grefenstette, John J.
PDF Url : ADA294103
Report Date : 12 AUG 1992
Pagination or Media Count : 12
Abstract : Truly autonomous vehicles will require both projective planning and reactive components in order to perform robustly. Projective components are needed for long-term planning and replanning where explicit reasoning about future states is required. Reactive components allow the system to always have some action available in real-time, and themselves can exhibit robust behavior, but lack the ability to explicitly reason about future states over a long time period. This work addresses the problem of creating reactive components for autonomous vehicles. Creating reactive behaviors (stimulus-response rules) is generally difficult, requiring the acquisition of much knowledge from domain experts, a problem referred to as the knowledge acquisition bottleneck. SAMUEL is a system that learns reactive behaviors for autonomous agents. SAMUEL learns these behaviors under simulation, automating the process of creating stimulus-response rules and therefore reducing the bottleneck. (AN)
Descriptors : *LEARNING MACHINES, *KNOWLEDGE BASED SYSTEMS, ALGORITHMS, COMPUTERIZED SIMULATION, SCENARIOS, REAL TIME, REASONING, TRACKING, RULE BASED SYSTEMS, REACTIVITIES, LONG RANGE(TIME), PROBLEM SOLVING, NAVIGATION, PLANNING, DECISION THEORY, DATA ACQUISITION, ADAPTIVE SYSTEMS, SELF OPERATION, SYSTEMS ANALYSIS, CONTROL THEORY, AUTOMATIC, HIGH LEVEL LANGUAGES.
Subject Categories : Cybernetics
Distribution Statement : APPROVED FOR PUBLIC RELEASE