Linear closed-form precoding of MIMO multiplexing systems in the presence of transmit correlation and Ricean channel

This paper presents a closed-form linear precoder for a MIMO spatial multiplexing (SM) system in the presence of transmit correlation and a Ricean component. Existing SM (V-BLAST and similar schemes), based upon channel matrix inversion, rely on the linear independence of antenna channel responses for stream separation and suffer considerably from high levels of fading correlation and/or dominating ill-conditioned line-of-sight channel components. We propose a simple algorithm that adjusts the transmitted constellation through power weighting and phase shifts that can be interpreted in some extreme case as a higher order constellation design scheme. We obtain a rate-preserving MIMO multiplexing scheme that can operate smoothly at any degree of transmit correlation and any type of LOS channel component. of the input's spatial signatures to detect the data, unless a form of joint encoding is applied across the streams to differentiate them. For example, current schemes like SM (e.g. V-BLAST) literally break down in the presence of correlation levels close to one or high Ricean factors. As a result, these algorithms simply fail to adapt themselves and extract the non-zero capacity that is present in highly correlated or strongly Ricean channels. Designing appropriate transmission techniques that can adjust to various kinds of channel and terrain scenarios is therefore an important and practical issue for the successful deployment of MIMO systems. Additionally, the transmit correlation and LOS components can be assumed to remain static over a longer period of time and parameters for these statistics can be fed back to the transmitter