Impact of Spatial Correlation in MIMO Radar

Abstract

The impact of spatial correlation on mutual information (MI) is analyzed for MIMO radar. Unlike the work done in literature for statistical MIMO radar, we consider the spatial correlation of the target matrix elements to study the correlated MIMO radar performance. There is a trade-off between coherent processing gain in correlated MIMO radar and spatial diversity gain of target scatterers in uncorrelated MIMO radar. We address how the MI between the received signal and target channel matrix is affected by spatial correlation. Using majorization theory and the notion of Schur-convexity, we prove that MI has a changing behavior with respect to spatial correlation, where at low SNR, the MI is Schur-convex, i.e. showing increasing performance as correlation increases. However, this behavior changes at high SNR, since MI is Schur-concave at high SNR, hence it decreases as the spatial correlation increases. Moreover, we investigate the conditions for spatially uncorrelated MIMO radar. According to these conditions, as the operating frequency increases with respect to the target location and dimensions, the received paths become more uncorrelated. Hence, the setup with lower operating frequency (more correlated) performs better compared to the higher frequency setup at low SNR. However at high SNR, this behavior is reversed.