无需信源数估计的宽带多径信号DOA估计算法

DOA Estimation Algorithm of Wideband Multipath Signals without Estimating the Number of Sources

现有的宽带DOA(direction-of-arrival)估计算法大都需要信源数已知,而实际应用中较难准确估计出信源的数目,信源数估计的误差将导致DOA估计的性能下降。针对此问题,文章提出一种无需信源数估计的宽带多径信号DOA估计算法。该算法首先利用泰勒级数展开将阵列接收数据近似为窄带数据,然后将宽带多径信号DOA估计问题转化为类似波束形成方法的约束最优化问题,最后通过求解最优权矢量得到最终的DOA估计值,并分析了约束参数的取值范围。仿真实验表明,相比于传统的TCT(two-sided correlation transformation)算法,文章算法无需信源数估计且估计性能下降不大。实测数据验证了文章算法的有效性。

The existing wideband DOA estimation requires the prior knowledge on the number of sources and there is a decline of estimation performance in case that the number of source is not correctly estimated which often happens in the practical application. To solve this problem,a novel DOA estimation algorithm of wideband multipath signal without estimating the number of sources is proposed. The received array signals are firstly transformed into narrow-band data by the Taylor series expansion,then DOA estimation is transformed into the framework of beamforming by introducing a new constrained optimization problem. Last,the optimal weight vector is utilized to get the final DOA estimation. Furthermore,the theoretical bounds of the constrained parameter are derived.Simulation results show that the proposed algorithm doesn＇t need to estimate the number of sources and has a little decline of estimation performance compared with the traditional TCT algorithm. The effectiveness of the proposed algorithm is demonstrated by the measured data.