Conventional linearly constrained adaptive beamformers often suffer from severe signal cancellation in the presence of interferers correlated with the signal. We propose a partially adaptive beamforming technique for correlated interference rejection in broadband signal environments. The beamformer output mean squared error is decomposed into an interference mean squared error term and an additional signal cancellation term that is due to the presence of correlated interference. Both mean squared errors depend on the adaptation space. The partially adaptive beamforming technique proposed here chooses an adaptation space which results in little signal cancellation while maintaining satisfactory interference cancellation. It is shown that, for a given interference scenario, a partially adaptive beamformer can be designed such that maximum interference cancellation is achieved without any signal cancellation. In practice, an approximate design procedure is provided to accommodate a set of likely interference scenarios. Analysis of the feasibility of this approach is presented. The effectiveness of the technique is demonstrated through examples.
展开▼
