Successive Interference Cancelation via Rank-Reduced Maximum Likelihood Detection

This paper proposes a codeword-based iterative detecting and decoding (IDD) algorithm using a rank- reduced maximum likelihood (ML) detector with pre- whitened interference over multiple-input multiple- output (MIMO) channels. This iterative algorithm operates on the principle of successive interference cancelation (SIC) where all the layers per codeword are decoded together at each iteration. The recent wireless standard requires to encode data bits per codeword, not per layer at multiple antennas. Thus, SIC algorithms based on layer-separating detectors could lose joint information between layers in a single codeword. Instead, the proposed algorithm minimizes this loss by using rank-reduced ML detectors over soft feedback. Simulations are performed on a space-time bit-interleaved coded modulation over Rayleigh fading channels, and demonstrate the proposed SIC algorithm is superior to comparable iterative and non-iterative IDD algorithms.