On-line parameter identification and control parameters optimization of robot joints with unknown load torque

Abstract

An online identification method based on discrete model reference adaptive identification is presented in this paper. First, the algorithm obtains the difference equation of the reference mode, which includes both load inertia and damping coefficient, by dispersing the mechanical movement equation of motion of motor. Landau's discrete time-recursive parameter identification is used to reduce the adaptation mechanism. Then, the control parameters are optimized on-line on the base of the estimation result and the error threshold. This method can estimate the load inertia and damping coefficient at the same time which are used to optimize the speed PI regulator parameters and improve the performance for speed control. The simulation and experiment results prove the effectiveness of the proposed method.