A NEURAL-NETWORK-BASED CLASSIFICATION OF ENVIRONMENT DYNAMICS MODELS FOR COMPLIANT CONTROL OF MANIPULATION ROBOTS

In this paper, a new method for selecting the appropriate compliance c ontrol parameters for robot machining tasks based on connectionist cla ssification of unknown dynamic environments, is proposed, The method c lassifies the type of environment by using multilayer perceptron, and then, determines the control parameters for compliance control using t he estimated characteristics. An important feature is that the process of pattern association can work in an on-line mode as a part of selec ted compliance control algorithm. Convergence process is improved by u sing evolutionary approach (genetic algorithms) in order to choose the optimal topology of the proposed multilayer perceptron. Compliant mot ion simulation experiments with robotic arm placed in contact with dyn amic environment, described by the stiffness model and by the general impedance model, have been performed in order to verify the proposed a pproach.