Study Of Robust Wideband Adaptive Beamforming

How to improve the robustness of adaptive beamforming is more and more important as its extensive use in radar, sonar, communication and seismic exploring. However, most of the early beamfomers are based on narrowband signal. Wideband signal beamforming is good for target detection, parameter estimation and target feature extraction because more frequency components in wideband signals can be synthetically used. So wideband signal beamforming is widely applied to acoustic imaging, underwater signal processing and speech acquisition and becomes an important research topic.Robust Capon beamforming techniques are studied and summarized first in this paper. The Robust Capon beamforming (RCB) proposed by Dr. Li in the University of Florida is studied in detail. The RCB has the robustness against array steering vector errors, finite sample size and excellent interference suppression capability. The constant beamwidth is a very important factor for the wideband signal beamformer. As an efficient wideband constant beamwidth beamforming method, the focused beamformer is not robust enough when the error of signal direction exists. To overcome this shortcoming, a robust broadband constant beamwidth beamforming method based on focusing approach (R-CBRCB) is proposed in this paper. The R-CBRCB mainly combines the idea of RCB with the traditional focusing approach and behaves the better robustness performance than the traditional focused beamformer. This new method also has some other advantages, such as computational efficiency, constant mainlobes, lower sidelobes, higher resolution and stronger interference rejection capability. It provides a promising approach to consistently estimate the SOI power for acoustic imaging. Simulation results show the feasibility of the proposed method.