利用整体退火遗传算法优化阵列改善方位估计

Performance Improvement in Estimating Direction-of-Arrival by Genetic Algorithm Combined with Simulated Annealing

该文将非比例选择、保证收敛且易于判断收敛的整体退火遗传算法应用于优化线阵方位估计性能。优化过程中,不仅优化阵元位置改善方位估计,而且将阵元个数作为优化变量,从而提供了更多的自由度来控制阵列性能。在满足空间谱估计精度的条件下,利用较少的阵元实现超分辨估计。优点是减少信号处理量的同时,简化了设备。仿真实验结果表明该方法收敛速度快,有极强的避免过早收敛及避免局部极值的全局优化能力。

A new type of genetic algorithm combined the simulated annealing with the advantage of guaranteeing convergence and deciding convergence easily is used for optimization the performance in estimation direction-of-arrival of linear array. In the process of optimization, not only the position of element, but also the number is used as variable. This can provide more degree of freedom to control characters of non-uniform linear array. It can use fewer elements to obtain the estimation accuracy of DOA. And it also can simplify equipment and decrease the computation complexity. The simulation results show that this method has strong ability to find the global optimization solution.