极大似然高分辨波达估计新方法

A novel algorithm for maximum likelihood high-resolution direction-of-arrival estimation

摘要: 极大似然估计器是波达方向估计中公认的最佳估计器,但是计算量很大。为了解决极大似然估计器由于进行多维格形搜索而带来的计算量大的不足,将粒子滤波方法与极大似然估计相结合,提出了一种基于粒子滤波的极大似然波达方向估计器(Maximum Likelihood DOA Estimator Based on Particle Filtering,简称MLE-PF)。研究结果表明,MLE-PF不但保持了原极大似然估计方法的优良性能,大大减小了计算量,计算复杂度由O(LK)降至O(K×Ns),而且在低信噪比时也具有比MUSIC以及MiniNorm方法更加优越的估计性能。

Abstract: A novel maximum likelihood high-resolution direction-of-arrival (DOA) estimator is proposed based on particle filtering. The new method,Maximum Likelihood DOA Estimator based on Particle Filtering (MLE-PF),is presented to deal with the computational load of a multidimensional grid search for Maximum Likelihood Estimator (MLE),which performs best among all the methods for DOA estimation. Simulation results show that MLE-PF keeps the perfect performance of MLE and lowers the computational complexity of MLE from O(L^K) to O(K×N_s). MLE-PF also performs better than MUSIC and Mi-niNorm,especially at low SNRs.