Adaptive Control Of A Legged Robot Using A Multi-Layer Connectionist Network

摘要

ABSTRACTA Multi -layer Connection Network (MCN) is to control a one -legged mobile robot. The networkhas no knowledge of the dynamics of the robot, and learns to develop a contol strategy through trialand error. Our results are presented in the form of computer simulations that demonstrate the ability of the (MCN) to devise a set of proper control signals that will develop stable running on a flat terrain. 1. INTRODUCTIONArtificial Neural Networks (ANN) have become a topic of intensive research, a fact that is manifested in the large number of models, algorithms, analysis studies and applications which have appeared in the last decade in the technical literature. The application of neural networks and neurocomputing to problems in robotics and control is a goal pursued by several research groups(e.g. Bavarians, Anderson2 and his references) There are several reasons for this interest. Viewed asa problem that seeks a solution, questions in adaptive control and in robot control can potentially beaddressed effectively by systems that can learn from example, learn from experience, and possessrobustness. Several neural network models have demonstrated such properties. Viewed as solutionsthat seek a problem, neural networks, which at present are ineffective as large -scale systems becauseof scaling problems, could potentially be effective in handling small -scale problems. The control of alegged robot with multiple degrees of freedom is a good candidate.