Design of Real-Valued Wideband Beamformers Using an Adaptive-Array-Theory-Inspired WLS: Theory and Algorithm

Abstract

Listen

Translate article

The weighted least squares (WLS)-based beamformer design can achieve flexible beampattern control via adjusting the WLS weighting function inspired by the adaptive array theory (AAT). Nevertheless, the weighting function is conventionally adjusted in an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ad hoc</i> way, and precise beampattern control cannot be guaranteed. Although an analytical solution for the weighting function adjustment was proposed very recently, it is only applicable to complex-valued narrowband beamformers. In some applications such as acoustic/seismic array processing, however, real-valued wideband beamformers are required, since acoustic/seismic signals are usually unmodulated, real-valued and wideband. This paper studies the theory and design of real-valued wideband beamformers using an AAT-inspired WLS, which can achieve precise wideband beampattern control. The main contributions of this paper include: i) we theoretically prove the existence of the WLS weighting function for the real-valued wideband beamformers to achieve precise control of wideband beampatterns, and derive the corresponding analytical solutions; ii) we provide the insights into the impact of the weighting function on the cost function, which is useful to find the optimum solution for the weighting function among multiple feasible solutions; and iii) we propose an algorithm for the design of real-valued wideband beamformers using the proposed AAT-inspired WLS, which does not require to fine-tune any user parameters, and can guarantee precise control of wideband beampatterns as demonstrated by the simulation results.