ON AN EFFECTIVE DESIGN APPROACH OF CARTESIAN SPACE NEURAL-NETWORK CONTROL FOR ROBOT MANIPULATORS

It is well known that computed torque robot control is subjected to pe rformance degradation due to uncertainties in robot model, and applica tion of neural network (NN) compensation techniques are promising. In this paper we examine the effectiveness of neural network (NN) as a co mpensator for the complex problem of Cartesian space control. In parti cular we examine the differences in system performance of accurate pos ition control when the same NN compensator is applied at different loc ations in the controller structure. It is found that using NN to modif y the reference trajectory to compensate for model uncertainties is mu ch more effective than the traditional approach of modifying control i nput or joint torque/force. To facilitate the analysis, new NN trainin g signal is introduced and used for all cases. The study is also exten ded to non-model based Cartesian control problems. Simulation results with three-link rotary robot are presented and performances of differe nt compensating locations are compared.