An Improved MIMO Transmission Diversity Smoothing Method by Constructing Cross-Covariance Matrices

In this letter, we propose a novel preprocessing method to improve the transmission diversity smoothing (TDS) effect. Firstly, by applying different matched-filters to the receiving data, we acquire a series of virtual subarray data with identical array manifold. Then, a batch of cross-covariance matrices are constructed with different subarray data. On this basis, multiplying all obtained cross-covariance matrices and auto-covariance matrices by their own conjugate transpose, we get a series of high-order covariance matrices. Afterwards, averaging all these new matrices, we reconstruct a smoothed data covariance matrix for bearing estimation. Compared with the existing TDS based methods, the proposed one makes improvement by introducing cross-covariance matrices to smoothing, rather than further increasing the number of auto-covariance matrices. Theory shows that the proposed scheme well reserves the decorrelation effect of TDS and improves the estimation accuracy by enhancing the array signal to noise ratio. Numerical simulations verify the effectivity and superiority of the modified TDS method with the comparison of root mean square error for bearing estimation of coherent targets.