An improved nonlinear adaptive inverse control systems based on filtered-ε LMS algorithm

A basic NAICS (nonlinear adaptive inverse control system) based on filtered-epsiv LMS algorithm has a complex structure. It is composed of more than ten adaptive nonlinear filters and four of them need to be trained on-line. Moreover, the disturbance canceller could not work well in the system. In order to simplify the structure and cancel the disturbance effectively of the system given above, an improved NAICS based on filtered-epsiv LMS algorithm is presented in this paper. It has a new disturbance canceller that candles the disturbances by a copy model of the left-inverse model of the nonlinear plant. Moreover, such disturbance canceller doesnpsilat need to train any filters. The improved system is composed of six adaptive nonlinear filters and three of them need to be trained on-line. Simulation results show the system can cancel the disturbance effectively and the inverse controller can converge quickly.