Improving MIMO detection by L-value analysis and adaptive threshold based cancellation

Abstract — We present an improved algorithm for detection of the received signal of a convolutionally encoded wireless MIMO system. The method is based on BLAST (Bell-Labs Layered Space-Time) algorithm with successive interference cancellation. Compared to existing solutions, cancellation is done adaptively. The receiver uses an MAP soft output demapper combined with an iterative (Turbo-) decoder. Adaptive symbol cancellation is done using a threshold which takes reliability of a symbol decision into account. Cancellations of wrongly decided symbols are reduced to a large extend. As a result, we obtain an SNR gain of 1 to 1.5 dB for Rayleigh fading MIMO channels with additive white Gaussian noise. I