Phy layer abstraction for SU-MIMO LTE system employing parallel interference-aware detection

Abstract

Maximum-likelihood (ML) receivers are optimum receivers for MIMO systems, but their complexity grows exponentially with the modulation order of the codeword and the number of spatial layers. The Reduced Complexity ML (R-ML) receivers based on low complexity adaptation of the max-log MAP detector and reduction of searching space are a good compromise to satisfy complexity-performance trade-off. In this paper we investigate physical layer abstraction for singleuser MIMO system with a R-ML receiver performing Parallel Interference Aware (PIA) detection. We develop a light-weight extended Mutual Information Effective SINR Mapping (MIESM) methodology based on look-up tables (LUT), which take the special non-linear nature of the PIA receiver into account. The results show that the proposed LUT-based MIESM abstraction method is as accurate as the direct MIESM abstraction, but at the same time has significantly less computational complexity. It can thus be easily used in performance prediction devices as well as system level simulators.