ROBUST ADAPTIVE BEAMFORMING BASED ON PARTICLE FILTER WITH NOISE UNKNOWN

Adaptive beamforming, which uses a weight vector to maximize the signal-to-interference-plus-noise ratio (SINR), is often sensitive to estimation error and uncertainty in the parameters, such as direction of arrival (DOA), steering vector and covariance matrix. Robust beamforming attempts to mitigate this sensitivity and diagonal loading in sample covariance matrix can improve the robustness. In this paper, beamformer based on particle filter (PF) is proposed to improve the robustness by optimizing the diagonal loading factor in sample covariance matrix. In the proposed approach, the level of diagonal loading is regarded as a group of particles and optimized using PF. In order to compute the postprobability of particles beyond the knowledge of noise, a simplified cost function is derived first. Then, a statistical approach is developed to decide the level of diagonal loading. Finally, simulations with several frequently encountered types of estimation error are conducted. Results show a better performance of the proposed beamformer than other typical beamformers using diagonal loading. In particular, the prominent advantage of the proposed approach is that it can perform well even noise and error in the steering vector are unknown