Soft comuting for autonomous robotic systems

Neural networks (NN), genetic algorithms (GA), and genetic programming (GP) are augmented with fuzzy logic-based schemes to enhance artificial intelli gence of automated systems. Such hybrid combinations exhibit added reasonin g, adaptation, and learning ability. In this expository article, three domi nant hybrid approaches to intelligent control are experimentally applied to address various robotic control issues which are currently under investiga tion at the NASA Center for Autonomous Control Engineering. The hybrid cont rollers consist of a hierarchical NN-fuzzy controller applied to a direct d rive motor, a GA-fuzzy hierarchical controller applied to position control of a flexible robot link, and a GP-fuzzy behavior based controller applied to a mobile robot navigation task. Various strong characteristics of each o f these hybrid combinations are discussed and utilized in these control arc hitectures. The NN-fuzzy architecture takes advantage of NN for handling co mplex data patterns, the GA-fuzzy architecture utilizes the ability of GA t o optimize parameters of membership functions for improved system response, and the GP-fuzzy architecture utilizes the symbolic manipulation capabilit y of GP to evolve fuzzy rule-sets. (C) 2000 Elsevier Science Ltd. All right s reserved.