Lattice Reduction Techniques in MIMO Systems with Correlated Channels

Multi-Input Multi-Output (MIMO) technology has become mature, and adopted in modern wireless systems. With multiple antennas at both transmitter and receiver, system data rate and throughput are greatly improved. Meanwhile, lattice reduction (LR)-aided detection has emerged as a promising scheme for signal detection in MIMO systems because of its low complexity and high performance. But in reality, channels between different antenna pairs are likely to be correlated, which impacts system capacity and performance. In this paper, the effect of spatial correlation on the performance and complexity of LR-aided detectors in MIMO systems is examined. Incorporating Kronecker model, first, we prove the diversity orders collected by LR-aided detectors over correlated MIMO channels, next, we measure their performance degradation through the parameter SNR gap when both transmit and receive correlation are present, and show that LR-aided detectors are more robust to spatial correlation compared to linear and SIC detectors, lastly, we study the effect of spatial correlation on the complexity of LR algorithm, and demonstrate through simulations that receive correlation affects LR-aided detectors more in terms of complexity compared to transmit correlation when MIMO size and correlation coefficient are large.