LR-Aided Detection for Large-MIMO System via Ant Colony Optimisation

In this work heuristic ant colony optimisation (ACO) procedure is deployed in conjunction with lattice reduction (LR) technique aiming to improve the performance-complexity tradeoff of detection schemes in MIMO communication. A hybrid LR-ACO MIMO detector using the linear minimum mean squared error (MMSE) criterion as initial guess is proposed and compared with other traditional (non)linear MIMO detectors, as well as heuristic MIMO detection approaches from the literature, in terms of both performance and complexity. Numerical results show that the proposed LR-ACO outperforms the conventional ACO MIMO detector, as well as the proposed ACO detector with the MMSE solution as initial guess, with a significant complexity reduction. The LR-ACO-based MIMO detector achi- eves a substantial improvement in terms of BER performance while expending lower computational resources than others ACO- based MIMO detectors considered. The performance-complexity results suggest that the proposed LR-ACO-MIMO is a promising solution for large number of antennas and/or modulation order, or even correlated MIMO channel scenarios.