Improved Detection Algorithm for Multiple-Input Multiple-Output Wireless Communication Systems

摘要

Maximum likelihood detection (MLD) algorithm provided the best bit error rate (BER) performance for MIMO communication system. Calculating complexity of MLD increased exponentially with the constellation size and the transmit antenna number which prohibited it to use in practical system.QR decomposition with an M-algorithm (QRD-M) kept M reliable candidates instead of deciding the signal. Decision was made after all detection layers processed. To accomplish near-MLD performance for QRD-M algorithm, M should be large enough for the selected paths to include the correct one.It resulted in the increasing of the complexity. If reducing candidate branch, the detection performance would become worse. Ordered successive interference cancellation (OSIC) algorithm with lower computing complexity would suffer from error propagation when an incorrect symbol was selectedin the early layers. To reduce these problems, an improved detection scheme was proposed in the paper which providing two adjustable parameters. Trade-off between detecting performance and calculating complexity could be obtained by setting the two parameters at different values. Simulationexperiments were given under the multiple scattering wireless communication environments, and the simulation experiment results showed the validity of proposed algorithm.