A Novel MIMO Signal Detection Method Using Hamiltonian Monte Carlo Approach

Multiple-input multiple-output (MIMO) is a modern wireless transmission technology that can dramatically increase communication speed and capacity. Since the computational complexity of received signal detection increases with the number of antennas, various improvement methods have been investigated. This paper focuses on Markov chain Monte Carlo methods and proposes a new signal detection method based on the Hamiltonian Monte Carlo. The proposed method can improve the signal search efficiency by intentionally expanding the discrete signal detection problem into a continuous-valued one. Simulation results show that the proposed method outperforms the conventional method using the Gibbs sampling method and achieves near-optimal performance, especially when the modulation order and spatial correlations are high, which are generally difficult for signal detection. This result suggests that the proposed method is a promising candidate for a practical MIMO system.