为了提高遗传算法(GA)控制的自适应光学(AO)系统的收敛性能,建立了一套新型的基于泽尼克(Zernike)模式系数的19单元自适应光学系统模型。在优化过程中,遗传算法不直接优化变形镜(DM)19个驱动器上的电压值,而是优化前10阶泽尼克模式系数。推导出19个电压值与前10阶泽尼克模式系数之间的关系矩阵,并进行了对比数值仿真。结果表明,该系统能够更好地校正固体激光器系统输出光束的波前像差。相对于直接优化变形镜电压值的无波前自适应光学系统,该自适应光学系统能够将遗传算法的收敛速度提高5倍以上。

To improve the convergence speed of adaptive optics (AO) system controlled by genetic algorithm (GA), a new 19-element AO system model based on GA is set up. The system chooses the first 10-order Zernike mode coefficients on the deformable mirror (DM) as the optimization function, rather than the 19 voltages on the DM. The transform matrix between the 19 voltages and the first 10-order Zernike mode coefficients was deduced and comparative mathematic simulations were done. The simulation shows that the correction ability of the proposed AO system is much better than that optimizing the 19 voltages. Moreover, compared with the GA optimizing DM voltages, the convergence speed of GA optimizing Zernike mode coefficients is improved by more than five times.