BACKPROPAGATION NEURAL NETWORKS FOR IDENTIFICATION AND CONTROL OF A DIRECT-DRIVE ROBOT

A neural approach is proposed to estimate parameters in dynamics of a direct drive robot. Before the estimation, the input-output data for i dentification are generated in a sequential and term-by-term manner fi rst. Then a two-layer neural network for parameter identification is p roposed, in which the back-propagation training method is used to adju st the weights between neurons. The goal is to find the weights that m inimize the root-mean-square error between the identification data and output of the network. With the estimated dynamics, existing trajecto ry-tracking algorithms, such as the well-known computed-torque method, can then be applied to make the robot move along a desired trajectory .