IIR Filter And MRAC Studay Of Linear Adaptive Inverse Control

Adaptive inverse control used for designing controller and modulator is first proposed by a famous professor Widrow in Standford University in 1986. Its principal is using the controlled object inverse model as a controller, the inverse model recognized by adaptive filter is series connected by object input, and it can control the object dynamic performance. Generally speaking, there are two adaptive filter, one is adaptive FIR filter, another is adaptive IIR filter. Compared with adaptive FIR filter, the coefficient of adaptive IIR filter is less, and its refrain from rash action is sooner. Therefore, adaptive IIR filter has get more and more attention for its low calculation and good performance. Combined adaptive inverse control and adaptive IIR filter, we researched noise reduction. Sereval adaptive IIR filter arithmetic and application is researched in this paper.First, the paper simply introduces adaptive inverse control method, basic algorithm of adaptive FIR filter and adaptive IIR filter. Then we propose some new algorithm of adaptive IIR filter.Adaptive IIR filter based on hyperstable theory is studied. The paper put forward a new adaptive recursive filtering identified algorithm combining hyperstable theory and adaptive inverse control, which can overcome the object disturbance. The algorithm conquers the limitation to fall into local particle of using the IIR filter expressed with the format of transfer function when uses gradient method and resolves the problem to prove the convergence of the parameters. Furthermore, it eliminates the disturbance when the model and the object are completely matching, and the discerned model is more exact and practical.This paper also research the cancelling noise method based on adaptive inverse control. Make use of adaptive inverse control cancelling noise, at thesame time to combine the hyperstable theory. A new method for designing discrete MRAC which can overcome the random disturbance and an adaptive model tracking control are separatly proposed. The method separates to control exportation and disturbance , it can raise respective function as soon as possible. The result of emulation proves the feasibility of this method.