基于随机并行优化算法的自适应光束全场整形方法探索研究

Because of the restriction by laser cavity and gain distribution, the intensity of laser beam can only have specific distributions. However, laser beams with desired intensity distributions are usually required in practical applications. An important solution to the problem is full-field beam shaping, which can not only change intensity distribution of the input beam, but also can compensate wave front of the output beam. We introduce the stochastic parallel optimization algorithm into full-field beam shaping. The full-field beam shaping based on stochastic parallel optimization algorithm, which can adaptively adjust the phase distributions of dual deformable mirrors to achieve the beam with desired optical field distribution. Compared with the conventional beam shaping technique, this technique, which has the advantage of eliminating the measurement of intensity and phase distributions of the input beam, can reshape the beam with dynamic phase and intensity variations. Based on the combination of theoretical analysis and experimental verification, the construction of metric function, the reduction of the interaction of adaptive near field beam shaping and adaptive wave front compensation of the output beam, the control mode of deformable mirrors would be studied in detail. Based on the theoretical and experimental analysis, we will establish the evaluation standard of the output beam after full-field beam shaping. Based on this evaluation standard, the influence of spatial resolution on full-field beam shaping will be studied in detail. In addition, the adaptive full-field beam shaping of input beam with unsymmetrical intensity distribution and the input beam with wavefront distortion will also be studied in detail. The evaluation of ultimate capacity of the full-field beam shaping system will be given in this project. The conclusion of the study can establish a foundation for the practical application of adaptive full-field beam shaping based on the stochastic parallel optimization algorithm.

激光器输出光束由于受到激光谐振腔结构、增益分布等因素限制而只能具有特定光场分布，然而在实际应用中往往需要激光光场满足某种分布以提高效率。采用光束全场整形方法改变光束光强分布，并有效补偿整形光束的相位是解决这个问题的重要途径。与传统光束全场整形方法相比，基于随机并行优化算法的自适应光束全场整形，通过采用双变形镜实时调整光束的相位来实现具有所需光场分布的光束。与传统光束全场整形方法相比，具有不需测量入射光束的光强和相位分布，可用于动态光场变化光束的整形等优点。本项目采用理论分析和实验验证相结合的方法，深入研究系统性能评价函数的构建、近场光强整形和整形后光束波前校正之间的耦合及其消除、变形镜的控制方式等问题，建立适合评价整形后光束性能的评价标准，讨论变形镜的空间分辨率对光束全场整形的影响，探索非对称光束和波前畸变光束的自适应全场整形，评估系统的极限整形能力，为该技术的实用化奠定基础。

激光器输出光束由于受到激光谐振腔结构、增益分布等因素限制而只能具有特定光场分布，然而在实际应用中往往需要激光光场满足某种分布以提高效率。特定的输出光场分布严重制约了激光应用领域的拓展和应用效能的发挥。采用光束全场整形方法改变光束光强分布，并有效补偿整形输出光束的相位是解决这个问题的重要途径。紧密结合光束全场整形的现实困难和未来需求，以及该项技术的国际最新发展动态，本项目提出基于随机并行优化算法的自适应光束全场整形，将基于随机并行优化算法的自适应光学技术应用于光束全场整形中，通过采用双变形镜实时调整光束的相位来实现具有所需光场分布的光束。与传统光束全场整形方法相比，该方法具有不需测量入射光束的光强和相位分布，可用于光场动态变化光束的整形，理论上可以将任意光强分布光束整形为所需光强分布光束等优点，并且可以有效拓展常规光束整形系统的设计方法。. 本项目采用理论分析和实验验证相结合的方法，建立了相对较为完善的基于随机并行梯度下降算法的自适应光束全场整形数学物理模型和数值仿真研究平台，搭建了基于双纯相位型液晶空间光调制器的实验研究系统，实现了基于随机并行梯度下降算法的自适应光束全场整形。系统分析了性能评价函数的构建，结合近场光强整形和整形后光束的波前校正，提出并实现了能够有效降低光强起伏、光轴抖动等影响的光束全场整形性能评价函数。深入研究了变形镜的空间分辨率对光束全场整形的影响，提出并实现了基于模式法控制的高空间分辨率变形镜近场光强整形和整形后光束波前相位校正，有效地提高了系统的整形精度和收敛速度。系统评估了近场光强整形和整形后光束波前校正之间的相互耦合，提出并对比研究了双变形镜同时控制用于光束全场整形的多种方案，探索出了最优控制方案，即整形变形镜采用光强整形性能评价函数，同时相位补偿变形镜采用相位校正性能评价函数，光强整形和相位补偿同步进行。对比研究了基于双非球面镜的光束全场整形和基于随机并行梯度下降算法的自适应光束全场整形，揭示了光束整形系统面形相位分布的生成规律，为系统性能的进一步优化提升提供了有效的指导。探索研究并实现了非对称光束和波前畸变光束的优化式自适应全场整形，拓展了光束全场整形方法在光束净化、远场焦斑光强调控和高清晰成像等领域中的应用。评估了系统的极限整形能力，为推进该技术的实用化奠定了坚实的基础。