Encoder position feedback based indirect integral method for motor parameter identification subject to asymmetric friction

Servo motors have been widely used in the automation industry for many years. Its control performance will directly affect the quality of the produced products. For the design of the servo motor controller, accurate modeling and parameter estimation will be one of the key design steps. Although a considerable amount of literature has discussed the modeling and parameter identification of servo motors, most of them focus on symmetrical friction parameter models and assume that motor velocity information is available. Based on the practical experimental examination, the displacement movements of the servo motor driven by harmonic input appear to be a drift phenomenon, which concludes that the friction force should be asymmetric. Moreover, coarse encoder quantization error during the practical measurement is also a problem that causes noisy velocity and acceleration estimations. These measurement imperfections would lead to inaccuracy of parameter identification results. In order to solve these issues, this paper presents an asymmetric friction model and an indirect integral method (IIM). The asymmetric friction model is able to capture the nonlinear position drifting phenomenon. For the proposed IIM, the use of velocity information is avoided. Moreover, an optimization algorithm is developed to minimize the quantized output prediction. Compared with the direct difference method (DDM) and the filtered regression model (FRM) in the existing literature, the numerical simulations, as well as the experimental validations, reveal that the proposed IIM has better parameter estimation performance than both the DDM and the FRM.