Generalised reduced-rank structure for broadband space-time GSC and its fast algorithm

Space-time generalised sidelobe canceller (GSC) is a well-substantiated method for broadband beamforming. In this paper, a generalised reduced-rank structure is established for the broadband space-time GSC. The generalised structure estimates signal subspace by a reduced-rank transform matrix, and so the dimension of broadband signal data vector is decreased. Therefore, the computational complexity for obtaining the adaptive weights is reduced. Generally, the reduced-rank matrix is obtained by eigen-decomposition of the data covariance matrix, which is a large computational load for broadband space-time processing. A fast reduced-rank algorithm is then proposed for the generalised structure to rapidly construct the reduced-rank matrix and to alleviate the computational burden caused by eigen-decomposition. The signal subspace is rather robust to the received data, hence, a set of received data can be used to construct the reduced-rank matrix for a rough and fast estimate of the signal subspace. With the undistorted response to signals of interest and the satisfactory anti-interference capability, the proposed algorithm could effectively reduce the computational complexity of the adaptive weight approach. Simulation results highlight the correctness and effectiveness of the proposed methods. HighlightsA new low-complexity space-time generalised sidelobe canceller (GSC) and its fast algorithm are proposed in this study.The GSC uses the reduced-rank matrix to estimate the signal subspace.The received data vectors are used to rapidly form the reduced-rank matrix.The singular value decomposition operation for the frequency domain constraints is applied to the novel algorithm.