Improved wavelet-based beamformers with dynamic subband selection

Adaptive beamformers, which can automatically adjust the weights to form prescribed spatial selectivity, have found applications in various facets of signal processing applications ranging from radar and sonar to wireless communications. To facilitate real-time implementations, it is of importance to mitigate the computational load in the weight adaptive process, especially for arrays which consist of many sensors in order to achieve better interference rejection and resolution. For this purpose, various attempts have been made for the design of partially adaptive beamformers, which only employ a subset of weights in the adaptive process. Despite their effectiveness, these approaches often call for computationally intensive eigendecomposition or nonlinear optimization. We address a simple, yet effective partially adaptive beamformer with the generalized sidelobe canceler (GSC) as the underlying structure. The new beamformer employs a set of P-regular, M-band wavelet filters in the design of the blocking matrix involved. The beamformer is applicable to the broadband scenario. Furthermore, a new subband selection scheme is addressed, which only retains the principal subband components in the adaptive process.