论文部分内容阅读
阐述了差分光柱像运动激光雷达探测大气湍流廓线的基本原理,针对雷达现有Levenberg-Marquardt反演算法对高空湍流反演误差大的问题,提出了带不等式路径约束的反演新模型,并采用惩罚函数法处理该模型,通过增加高空湍流信息量,避免了非物理意义的反演解;同时为了减弱现有算法对初值和先验知识的依赖,进一步提出了基于遗传算法的初值寻优策略,能够将现有算法的初值定位在全局空间内。利用改进算法和现有算法数值仿真了典型的大气湍流廓线,并对合肥地区实测激光雷达数据进行了分析。结果表明,改进算法增强了迭代过程的全局搜索能力,对测量误差有较强的稳健性,能够有效提高反演精度和高空湍流的准确性,同时也加快了收敛速度。
In this paper, the basic principle of differential beam-moving lidar for detection of atmospheric turbulence profile is described. Aiming at the large inversion error of the existing Levenberg-Marquardt inversion algorithm for high altitude turbulence, a new inversion model with inequality path constraints is proposed The penalty function method is used to deal with this model. By increasing the information amount of the turbulence at high altitude, the non-physical inversion solution is avoided. At the same time, in order to reduce the dependence of the existing algorithm on the initial value and prior knowledge, the initial value based on genetic algorithm Optimization strategy, the initial value of the existing algorithm can be located in the global space. The typical atmospheric turbulence profile is simulated by using the improved algorithm and the existing algorithm. The data of laser radar in Hefei area are analyzed. The results show that the improved algorithm enhances the global search ability of the iterative process and has strong robustness to measurement error, which can effectively improve the accuracy of inversion and the accuracy of high-altitude turbulence and speed up the convergence speed.