Experimental Identification of DC Electric Drives

Abstract

Because of increasing joint speed and acceleration of complex mechanical structures as rigid and flexible robot manipulators, accurate models of joints with electric drives are needed to improve their simulation and their control. We propose a new way to identify the electrical and mechanical parameters of a robot joint driven by a DC motor. The method is based on a closed loop identification of the inverse model which is linear in relation to the parameters. Least Squares techniques and exciting trajectories of current and velocity are used These trajectories minimize a criterion which is a function of the condition number of an observation matrix. An experimental identification of one prismatic joint EMPS300 driven by a DC permanent magnet motor is given.