利用改进遗传算法的DOA估计

Doa estimation using improved genetic algorithm

用极大似然估计 (MLE)得到到达信号的方向 (DOA) ,在统计性能方面要比其它一些理论优越 ,但是由于该方法为一种多维参数估计 ,采用常规搜索方法 ,精度受到网格限制 ,不能任意逼近最优解 ,并且容易收敛到局部最优。而遗传算法是一种有导向的随机搜索方法 ,它具有适用条件宽松 ,有较大的概率收敛到全局最优等优点。在此通过改进的遗传算法 (IGA) ,较好地解决了一般搜索算法存在的不足 ,计算机模拟实验证明其可行。

The maximum likelihood estimation (MLE) of DOA is an appealing algorithm. With the best Statistical Performance. However, with its multi dimensional parameter estimation computation load, the conventional searching method was limited by the minimum search step, and the ability of the conventional search to converge to global optimality is determined by the initial search point. Improved genetic algorithm (IGA) is an orienteed random search method, which has a moderate precondition, a good probability converge to global optimum. We resolved this problem with IGA, and convinced its viability with computer simulations.