Research On Correction Technology Of Wavefront Distortion From Atmospheric Turbulence By Adaptive Optics System

When the light beam propagates in the atmosphere,the atmospheric turbulence causes random disturbance to the phase of light field,which distorts the wavefront surface and causes the image quality of the optical receiving system to decline.The phase screen is an important means to describe the random disturbance.The adaptive optics system corrects the wavefront distortion in real time,so that the imaging system can overcome the interference of atmospheric turbulence and maintain good image quality.From the point of view of control theory,adaptive optics system is an automatic control system with wavefront phase as the controlled variable and wavefront corrector driving voltage as the control variable.The selection of control method is of great significance to improve the system performance.In this dissertation,the numerical simulation method of disturbance from atmospheric turbulence on wavefront phase and the control method of adaptive optics system are studied.Research results are shown as follows:(1)Aiming at problems of different methods for generating atmospheric turbulence phase screens,containing the absence of different frequency components,creasing problem and the large error of phase structure function,this dissertation proposes a new method named intrinsic embedding fractional Brownian motion for generating non-Kolmogorov atmospheric turbulence phase screen.According to the structure function and correlation function of non-Kolmogorov atmospheric turbulence phase screen,the relationship between it and fractional Brownian motion surface is analyzed.First,the phase covariance matrix satisfying the statistical law of atmospheric turbulence is constructed,which can be extended to a block circulant matrix with circulant blocks(BCCB matrix)with nonnegative eigenvalues;Next,the Cholesky decomposition of BCCB matrix is obtained by solving fast Fourier transform;Then,the complex Gaussian random matrix filter is used to generate the stationary Gaussian surface;Finally,the desired phase screen is obtained by a transformation from stationary Gaussian surface to fractional Brownian motion surface.Simulation results show that compared with existing methods including power spectral density,power spectral density-subharmonic,Zernike polynomials and random midpoint displacement method,the low and high frequency components of the phase screen generated by proposed method are both sufficient,and there is no creasing problem;Under the condition of different phase screen parameters,the deviation between the phase structure function and the theoretical curve is generally less than those of other current methods;Meanwhile,the time consumption of generating phase screens is significantly reduced when the proposed method is used.(2)To solve problems including low accuracy,high requirement for modeling accuracy and difficult implementation of different control methods for adaptive optics system with wavefront sensor,this dissertation proposes linear quadratic H? control method for adaptive optics system with wavefront sensor.The control method consists of H? state variable estimator and state feedback optimal control law.The H? state variable estimator takes the gain coefficient between the system noise and the estimation error of state variable as the cost function,and solves the state variable estimation value which makes the maximum gain coefficient reach the minimum value according to the optimization theory.The H? state variable estimator guarantees the upper bound of state variable estimation error and is time invariant,which reduces the amount of computation.The state feedback optimal control law takes the sum of steady-state error,dynamic error and energy consumption of the system as the objective function,and solves the control voltage to make it reach the minimum value.Simulation results show that compared to the current methods including direct slope method and linear quadratic Gaussian control,the times of diffraction limit,root mean square error of wavefront phase,Strehl ratio of facula,Zernike expansion coefficients,settling time,peak signal-to-noise ratio(PSNR)and structural similarity(SSIM)of extended target are all the best when the proposed method is used.Meanwhile,the correction effect can remain stable over time.(3)Aiming at problems of slow convergence speed and easily affected modeling accuracy of two kinds of control methods for adaptive optics system without wavefront detection,this dissertation proposes fast model-free control method for wavefront sensorless adaptive optics system.The Zernike expansion coefficient vector of wavefront phase correction is taken as the vector to be optimized.According to the statistical law of non-Kolmogorov atmospheric turbulence,different random disturbance amplitudes are set for every component of the vector to be optimized,and the signal of every component is determined by Hadamard matrix,Then,the vector to be optimized is transformed into the control voltage vector,and the vector to be optimized which makes the Strehl ratio of imaging facula reach the maximum value is searched iteratively.Simulation results show that compared to the current methods including simulated annealing,stochastic parallel gradient descent and Booth's method,the proposed method is the best when indexes including times of diffraction limit,root mean square error of wavefront phase,Strehl ratio of facula,Zernike expansion coefficients,settling time,PSNR and SSIM of extended target are considered comprehensively.Meanwhile,the correction effect can remain stable over time.(4)On the adaptive optics platform consisting of main control computer,high voltage driver module,piezoelectric deformable mirror,interfacing telescope with beamsplitter and wavefront sensor,response matrix calibration of adaptive optics system and proportional-integral-differential control is programmed.Experimental results show that the adaptive optics system makes the root mean square error of distorted wavefront surface reduce obviously,and energy focusability and Strehl ratio of imaging facula increase definitely.