【摘要】 基于特征分解理论的子空间类高分辨方位估计方法是目前阵列信号处理领域的研究重点。与波束形成法相比,子空间类法的特点为估计精度高和分辨能力强。但是这类高分辨方法对阵列模型失配十分敏感,存在阵列误差时其估计性能明显下降。文章通过仿真和实验深入研究了MUSIC、Johnson和Mini-Norm等子空间类高分辨方位估计方法的稳健性问题,分析了一定信噪比条件下阵列误差对上述三种方法估计结果的影响程度。研究结果表明,MUSIC法和Johnson法的估计性能相当,而Mini-Norm法的稳健性明显高于MUSIC法和Johnson法,分辨能力和估计精度较好,具有良好的工程应用前景。更多还原

【Abstract】 Subspace high-resolution direction-of-arrival (DOA) estimation algorithms based on eigen-decomposition theory have become a focus in array signal processing. These algorithms show good performance in resolution and precision compared to the beam-forming method. However, they are found to be very sensitive to array modeling errors. Their performances decrease drastically in the presence of array errors. In this paper, robustness of high-resolution direction-of-arrival estimation algorithms is studied with simulations and experiments. The effect of array errors on MUSIC, Johnson and mini - norm algorithms with certain signal-to-noise ratio (SNR) is analyzed. It is shown that, the performances of MUSIC and Johnson are quite similar, while mini-norm is more robust than the other two. This means that mini-norm has good performance in resolution and precision, therefore a good prospect in engineering applications.更多还原