Breadth-First Search Detection With Permutation Channel for Massive MIMO

The ubiquitous connectivity demands of the next-generation mobile networks pose challenges to multiple-input multiple-output (MIMO) detection with a great number of antennas and large alphabets. Traditional linear detectors struggle to meet the stringent performance requirements in dense MIMO configurations. While approaching maximum likelihood (ML) performance, breadth-first search detection (BFSD) requires numerous visited nodes (VNs) to maintain performance, incurring increased complexity and latency primarily distributed by metric calculation and path sorting of VNs. To maintain the performance with reduced complexity and latency, we propose a breadth-first search detection with permutation channel (PermBFSD) that employs varied symbol detection orderings by applying random permutations to the channel matrix. Moreover, we introduce early termination strategies that halt the search upon reaching a fixed or adaptive layer. The varied symbol detection orderings allow for early termination with all symbols estimated, thereby reducing complexity and latency. Simulation results and complexity analysis demonstrate that the proposed PermBFSD outperforms the prior arts with an up to 1.3 dB performance gain at the bit error rate (BER) of $10^{-4}$ and an 84.7% reduction in sorting complexity in the $16\times 16$ MIMO system.