Low-complexity detection and decoding based on shortest path search and lattice-reduction for MIMO systems

Abstract

In this thesis, we propose several low-complexity detection and decoding algorithms for multiple-input multiple-output (MIMO) systems. First, we propose a novel maximum likelihood (ML) symbol detection algorithm, which searches lattice points based on the shortest path algorithm. To reduce the computational complexity, we apply scaling, regularization, and lattice-reduction (LR) techniques to the proposed algorithm. Next, we derive the optimal LR-aided successive interference cancellation algorithm, which updates the mean and covariance of the effective symbol vector at each detection stage. Finally, we propose a LR-aided decoding algorithm for coded MIMO systems. The proposed decoding algorithm nearly achieves the performance of optimal iterative detection and decoding (IDD) with a maximum a-posteriori (MAP) detector even if it has much lower computational complexity.