一种改进均方误差性能的新型鲁棒性波束形成算法

A new robust adaptive beamforming with improved mean-squared error performance

结合信干噪比最大化和均方误差最小化两个优化目标,提出一种新型的鲁棒性波束形成算法。该方法考虑信号估计误差,在传统的最小方差的代价函数中引入信号协方差矩阵的估计误差,并在波达角估计误差的约束下,将鲁棒性波束形成器转换成基于支持向量机形式的波束形成器,通过一种高效的新型支持向量机训练算法计算阵列权值;然后以均方误差最小化为目标来修正阵列权值。仿真结果表明:该方法降低了波束形成器对信号估计误差的敏感度,提高了其抑制非平稳干扰的能力,且具有更好的均方误差性能。

This paper presents a new kind of robust beamforming method that provides joint improvement of SINR and MSE. We generalize the conventional linearly constrained minimum variance cost function by including the error matrix of signal covariance matrix and error constraints of DOA. The final cost function adopts the form of a support vector machine （SVM） for regression. To compute the beamformer weights, we adopt a computationally efficient Learning Algorithm for a new Regression SVM. Then we choose coefficient of beamformer vector to minimize the MSE. Computer simulations demonstrate an improved performance in comparison with other robust beamforming techniques.