Robust Adaptive Beamforming Based on Non-convex Quadratic Optimization with Semidefinite Relaxation
Yong Chen, Fang Wang · 2018
In practical application scenarios, the performance of robust adaptive beam forming depends on the accurate estimation of steering vector. However, the optimization problem that implements the steering vector estimation is usually non convex. In this paper, the relaxation technique is introduced into adaptive beam forming to solve the non-convex optimization problem and estimate the steering vector accurately. In addition, considering the widespread problem of signal self-cancellation in beam forming, the desired signal component is removed from the sample covariance matrix as far as possible. Computer simulation results show that the proposed SDR-RAB method is effective in steering vector mismatch scenarios, and its performance is better than the existing methods.