Reduced‐dimension MVDR beamformer based on sub‐array optimization

Abstract

A reduced-dimension minimum variance distortionless response based on sub-array optimization beamformer (SAMVDR) is proposed for large array systems or some hardware performance constrained scenarios. The SAMVDR allows to weigh between convergence rate and computational complexity by selecting the sizes of the sub-arrays. It is worth noting that, employing a sub-array strategy, the SAMVDR allows for computing the matrix inversion with flexibly choosing the size of the matrix. When the size of each sub-array is 1, namely single weight optimization MVDR (SWMVDR), the SAMVDR avoids matrix inversion. So the SWMVDR is easiest to implement in system than other sizes of sub-arrays in the SAMVDR. It is found that the SAMVDR has manageable computational complexity and fast convergence rate, which is verified by theoretical analysis and simulation results. Importantly, the output performance of the SAMVDR is very close to auxiliary vector (AV) scheme, and the SWMVDR has lower computational cost than the AV. In addition, the SAMVDR can be applied to parallel systems with higher tracking speed.