Research On Adaptive Optics System Control Technology Based On Model Identification

Adaptive optics technology can compensate the random wavefront distortion caused by the transmission medium in the transmission process in real time,and is widely used in astronomical observation,space target observation and laser transmission systems.In recent years,with the development of adaptive optics systems,the application of adaptive optics system has entered many fields such as optical communication,inertial confinement fusion,medical and medical,virtual reality,and holographic technology.As one of the key technologies of adaptive optics system,wavefront control technology directly affects the correction performance of adaptive optics system.At present,most adaptive optics systems use simple and easy to implement proportional integral control,but the parameter adjustment of this method requires human experience,and when adjusting parameters,the correction performance of adaptive optics system is often sacrificed to ensure the stability of the adaptive optics system.Although many control algorithms are proposed in the adaptive optics system,such as robust control,adaptive control,neural network,etc.al,most of them are stuck in the theoretical simulation stage,which is difficult to practically apply.Therefore,the existing control algorithms of adaptive optics systems have been unable to meet the increasing correction requirements of adaptive optics systems,which also limit the development of adaptive optics systems and applications in some emerging fields.In order to solve the control problem of adaptive optics,this paper proposes to use linear quadratic Gaussian control to improve the correction performance of adaptive optics system.For the first,since the linear quadratic Gaussian control requires an a model of the controlled object with high precision in application,in order to ensure that the linear quadratic Gaussian control can be realized in the adaptive optics system.According to actual adaptive optics system,a subspace identification method based on the errors-in-variables model is proposed.The state-space model of the adaptive optics system is established by using the input and output data of the adaptive optics system.According to the simulation results,the state-space model of the established adaptive optics system has higher accuracy and certain noise suppression ability and robustness.What is more,this approach provides a model basis for other model-based control algorithms.Secondly,according to the the state space identification model of adaptive optics system,the method of linear quadratic Gaussian control technology based on state adjustmen is adopted in this paper.The residual wavefront is minimized as the quadratic criterion,and the gain of the feedback control law is determined by minimizing the quadratic criterion.The state vector of the adaptive optics system is defined according to the Zernike polynomial expansion form of the incident wavefront and the linear relationship between the deformed mirror and the wavefront sensor.For the unknown state of the initial state of the adaptive optics system,the state vector of the adaptive optics system is filtered or estimated by the Kalman filter and the Kalman filter state.The linear quadratic Gaussian closed-loop control of the adaptive optics system can be achieved by solving the state estimation and the Kalman filter gain,and by minimizing the feedback gain obtained by solving the quadratic criterion.Then,the linear quadratic Gaussian control static wavefront and dynamic wavefront correction experiments are performed on the adaptive optics system,and compared with the proportional-integral control.The experimental results show that compaered with the proportional-integral control,the linear quadratic Gaussian control exhibits its unique advantages in terms of various performance indicators such as the peak intensity of far-field spot after correction.,the root mean square value of the residual wavefront,the convergence speed of the deformed mirror voltage,the power in the backet,and the control bandwidth,spot jitter suppression,system stability,robustness and so onFinally,the paper also verified the influence of adaptive optics system noise and Gaussian white noise on the wavefront correction effect and capability of linear quadratic Gaussian control of adaptive optics.From the experimental results,the approximate noise of adaptive optics system is used as the measurement noise.The linear quadratic Gaussian control has a significant improvement in the correction effect.This shows that the system noise has a great influence on the linear quadratic Gaussian control.If the noise model of the adaptive optics system can be accurately determined,it is possible to further improve the wavefront correction capability of the linear quadratic Gaussian control in the adaptive optics system.