Broadband MIMO Beamforming using Spatial-Temporal Filters and Polynomial Matrix Decomposition

For a narrowband MIMO system the SVD is shown to design a channel with the highest gain. This work is then extended to a broadband MIMO scenario where SBR2, a novel algorithm for calculating the EVD of a polynomial matrix is applied. The algorithm's operation, based on Jacobi rotations is discussed and a multi-channel beamformer is designed. This is evaluated against a transmitter antenna selection (TAS) and a transmitter-receiver antenna selection scheme (TRAS) in terms of channel gain. By using the dominant SBR2 mode for transmission and a matched filter to coherently combine the received the signals, a single very high gain wireless channel can be designed. Alternatively SBR2 may be applied at both the transmitter and receiver to design multiple orthogonal high gain wireless channels. The proposed SBR2 technique therefore has the potential to enhance spectral efficiency by exploiting spatial and temporal diversity in a wireless system.