On-line parameter tuning of disturbance compensation in precision positioning

This paper presents an on-line parameter tuning for the disturbance model using a least square method base on estimated disturbance errors. In this research, both nonlinear frictions and modeling errors between the mathematical model and the actual plant system are handled as disturbances in mechanism. It is well-known that the disturbance variation deteriorates positioning performances. A viscous friction and a motor torque constant are taken up a problem as primary factors in disturbance variation, because those parameters are frequently varied for temperature change due to drive conditions, such as before/after warm up motion. In the proposed parameter tuning, the disturbance variation is handled as an additive perturbation for the nominal disturbance, and a least square method for the additive perturbation estimates disturbance model parameter errors. The proposed positioning control approach with an adaptive disturbance model-based feedforward compensation has been verified by experiments using a table drive system on machine stand.