GENETIC-ALGORITHM-BASED FUZZY CONTROL OF SPACECRAFT AUTONOMOUS RENDEZVOUS

The combination of the control capabilities of fuzzy logic with the le arning capabilities of genetic algorithms is investigated. Fuzzy logic allows for the uncertainty inherent in most control problems to be in corporated into conventional expert systems. Although fuzzy-logic-base d expert systems have been used successfully for controlling a number of physical systems, the tasks of selecting acceptable fuzzy membershi p functions and rule sets have generally been accomplished via subject ive decision-making. In this paper, high-performance fuzzy membership functions and efficient rules for a fuzzy logic controller that manipu lates a mathematical model simulating the autonomous rendezvous of spa cecraft are discovered using a genetic algorithm, a search technique b ased on the mechanics of natural genetics. The membership functions an d rules discovered by the genetic algorithm provide for a more efficie nt fuzzy logic controller than membership functions selected by the au thors for the rendezvous problem. Thus, genetic algorithms are potenti ally an effective and structured approach for designing fuzzy systems. (C) 1997 Elsevier Science Ltd. All rights reserved.