Research On Hysteresis Modeling And Compensation Of Fast Steering Mirrors Based On BP Neural Networks

Due to the short communication wavelength,the satellite laser communication technology has the advantages of long communication distance,good confidentiality,strong anti-jamming ability and large transmission capacity,etc.,and has become a research hotspot of space communication technology in the world.The fast steering mirror is a very important driving device of the precision tracking system,the control of the fast steering mirror directly determines the acquisition and tracking performance of the precision tracking system.A fast steering mirror generally contains four ceramics that distributed in a crisscross pattern.The cooperative movement of each two ceramics along one radial direction generates the deflection of the fast steering mirror in the same orientation.Unlike the hysteresis nonlinearity of a single piezoelectric ceramic,which is symmetric or asymmetric,the fast steering mirror exhibits complex hysteresis characteristics.Based on the internal structure,the complex hysteresis characteristics of the fast steering mirror are analyzed,and three methods of hysteresis modeling are presented.Firstly,the complex hysteresis effect of the fast steering mirror is ignored,and the hysteresis model based on Madelung's rules is proposed.The information of the major ascending curve is used to describe the inner hysteresis curves.This method mainly has two aspects: one is the wiping-out mechanism;the other is that the as-cending and descending curves of the hysteresis are described by polynomials.The wiping-out and congruency properties of the proposed are investigated.Secondly,a structural hysteresis modeling method for fast steering mirror is presented.The input and output data of the fast steering mirror are processed,and the hysteresis modeling of a single piezoelectric ceramic is completed by PI model and nonlinear function.Then,the output of the fast steering mirror is deduced according to its internal structure.In addition,the method of parameter identification are proposed and characteristics of structural hysteresis modeling are analyzed by simulation.Thirdly,the neural network is used to model the hysteresis of the fast steering mirror.This hysteresis modeling method consists of symmetric hysteresis operator based on Madelung's rules and neural network.The hysteresis operator provides a basic hysteresis motion for the fast steering mirror and is used to expand the input space of the neural network.The neural network modifies the basic hysteresis motion to accurately describe the hysteresis nonlinearity of the fast steering mirror.The wiping-out and congruency properties of the presented model are showed by simulation.Finally,an experimental platform of fast steering mirror was built.Three methods were used for modeling,respectively.The experimental results verified the effectiveness of the presented methods.In order to compensate the hysteresis effect of the fast steering mirror and make the system approximate to a linear system,the hysteresis modeling method based on neural network was chosen to obtain the inverse model.