CONTROL OF SWITCHED RELUCTANCE MOTOR TORQUE FOR FORCE CONTROL APPLICATIONS

The paper presents a method for controlling Switched Reluctance Motor (SRM)'s torque for force control applications. SRM's are used in AdeptOne robots, and we perform experiments with two robots, controlled in coordination, in grasping and manipulation of various objects. The object and robot parameters are not exactly known, and adaptive methods are used to control the overall system. These methods are model-based control techniques which require high bandwidth torque control. This requirement is typical for high precision mechanisms. SRM's characteristics are very nonlinear. In particular the torque ripple, friction, and the torque versus position and current relationships were analyzed in the context above-mentioned, and specifically, for force control applications. The proposed method is based on a new commutation algorithm and a measured torque versus position and current relationship, used to smoothen the SRM's torque ripple, hence generating a torque output nearly independent of position. Furthermore, the internal friction is estimated on-line, and compensated for. This renders a high accuracy torque tracking. The torque control method is based on feedback from the motor angular velocity, motor angle, armature current, and feedforward for friction compensation and cancellation of nonlinear effects. The method has been tested experimentally on Adept motors and the results were very encouraging. The method has been also used for adaptive control of two coordinated Adept robots.