A survey of robotic control systems constructed by using evolutionary computations

Abstract

Provides a brief discussion on control and motion planning approaches to robotic manipulators or mobile robots with soft computing techniques. In particular, the paper presents the latest accomplishments and explores future trends in such a field. Specific topics of interest to this survey include: (1) the identification of a dynamic model for multi-link manipulators by using neural networks (NNs) with genetic algorithms (GAs), (2) behavior-based control of mobile robots or manipulators by applying NNs or fuzzy reasoning, together with evolutionary computations such as GAs, genetic programming (GP), or evolutionary strategy (ES), (3) a path planning problem with obstacle avoidance for mobile robots or manipulators by using evolutionary computations, and (4) a motion planning approach to configure self-organizing robots by applying GAs.