Inertia and Friction Estimation of a Velocity-Controlled Servo Using Position Measurements

Abstract

This paper proposes a method that estimates the parameters of a velocity-controlled servo. A proportional-integral controller, which uses only position measurements, closes the loop. The proposed approach uses the steady-state response produced by steps and sine-wave signals; they do not produce high levels of vibration on the servo compared with random signals commonly used with the least squares algorithm; moreover, it relies on simple numerical calculations. The method, which is called in the sequel as the steady-state response method (SSRM), consists of two steps. The first step uses three constant reference inputs in order to identify a constant disturbance and the viscous and Coulomb friction coefficients of the servo. In the second step, the SSRM estimates the servo inertia using a sine wave plus a constant signal as a velocity reference input and employs the estimate of the viscous friction coefficient obtained in the first step. Experiments on a testbed employing a brushless servomotor allow comparing the results obtained using the SSRM and those produced by a standard recursive least squares method (RLSM).