一种针对直接定位的联合搜索优化算法

Joint Search Optimization Algorithm for Direct Position Determination

针对传统网格搜索方法用于直接定位时计算量过大的问题,提出一种联合多种搜索方法的优化算法。首先利用DOA测向交叉定位确定一个粗略的搜索范围,然后在该范围中采用多重搜索算法,并在每重搜索中嵌套爬山算法,通过求解包含目标位置信息的共轭矩阵的最大特征值,得到相应的位置估计;此外,文章还分析了新算法的时间复杂度。理论分析和仿真实验可以表明,该算法的搜索准确率接近传统网格算法,能够求得逼近最小均方误差估计量的全局最优解,同时算法的计算量大幅减少。

An optimization algorithm combining multiple search methods is proposed to address the problem that calculating quantity of the traditional grid searching method is excessive when used for direct position determination.The DOA crossing location is employed to determine a rough search area.Then we use multiple-searching in the area and nest hill-climbing searching in every search process.The location is finally estimated by determining the maximum eigenvalue of Hermitian matrix containing target location information.In addition,the time complexity of the new algorithm is also analyzed.Theoretical analysis and simulation experiment show that the search accuracy of the new algorithm is close to traditional grid algorithm.And we can get global optimal solution which is close enough to minimum mean square error estimate.Meanwhile,the new algorithm can significantly reduce the amount of calculation.