摘要 该文针对短采样宽带信号近似最大似然(AML)方位估计计算量大的问题,将马尔可夫蒙特卡罗方法与近似最大似然方位估计相结合,提出一种基于Gibbs抽样的近似最大似然方位估计新方法(Approximated Maximum Likelihood DOA estimator based on Gibbs Sampling, AMLGS)。研究结果表明,AMLGS方法不但保持了原近似最大似然方位估计方法的优良性能,而且显著减小了计算量。把原方法的计算复杂度从O(LK)减少到O(K×J×Ns)。
Abstract:Approximated Maximum Likelihood (AML) estimator has been shown to be the best performance in short sampling wideband sources DOA estimation. However, the computation burden of AML is very large. In order to resolve the question of computation burden, Markov Monte Carlo methods are combined with Approximated Maximum Likelihood DOA estimator. A novel Approximated Maximum Likelihood DOA Estimator based on Gibbs Sampling (AMLGS) is proposed. AMLGS not only keeps the excellent performance of the original AML, but also reduces the computation greatly, from the computational complexity O(LK) of original method to O(K×J×Ns).