A novel nonlinear adaptive filter for modeling of rate-dependent hysteresis in Giant Magnetostrictive Actuators

Hysteresis is a nonlinear phenomenon, which exists in smart actuators, is usually undesirable and influences the performance of actuator. Therefore, it is necessary to establish a precise hysteresis model to eliminate hysteresis effects. A novel adaptive filter is proposed to model the rate-dependent hysteresis nonlinearity in a Giant Magnetostrictive Actuator. In the proposed filter, Generalized Play Operators are combined with linear delayed adaptive transversal filter to compose a new serial structure of adaptive filter model. Variable step LMS algorithm is used to adjust the weights value. In the end, the proposed adaptive filter is applied to model the rate dependent hysteresis of Giant Magnetostrictive Actuator. Experimental results show that the proposed Generalized Play Operator adaptive filter can describe the rate-dependent hysteresis behaviors including different single frequency input signal and multi-frequency composite input signal. The modeling results are agreed with experimental data well, the root mean square error is less than 0.5073 micron and the relative error is less than 3%.