Position Domain Iterative Learning Feedforward Control for Speed Fluctuation Suppression in Gimbal Servo System With Harmonic Drive

This study proposes a novel position-domain iterative learning feedforward control (PDILFC) method to suppress the output speed fluctuations, which are mainly induced by the nonlinear kinematic error of harmonic drive in the gimbal servo system. Considering the periodic characteristics of the kinematic error and output speed error, position domain control and feedforward control are employed. The stability and convergence conditions of PDILFC method in the position domain are derived by means of the domain transformation method. Meanwhile, to avoid the problems of random noise accumulation and resource waste due to multiple iterations, a filter with the forgetting factor and an iterative termination criterion are put forward, respectively. Then, the compensation table can be constructed and stored in the memory chips, and the iterative learning method no longer needs to be run during the operation of the gimbal servo system. Finally, the effectiveness and feasibility of the PDILFC are verified by numerical simulations and experiments on a control moment gyroscope. The results prove that the proposed method has good disturbance suppression on the speed fluctuations, and its accuracy can be improved by about 30%.