Low complexity signal detection employing multi-stream constrained search for MIMO communications

As a signal detection method for multiple-input multiple-output (MIMO) communications, this paper proposes multi-stream constrained search (MSCS) that achieves very good trade-off between computational complexity and bit error rate (BER) performance. The proposed method sets a minimum mean-squared error (MMSE) detection result to the starting point. From this point, MSCS searches for signal candidates in multi-dimensions of the noise enhancement from which the MMSE detection suffers. In the search, some streams of the signal candidates are fixed at constellation points. Among the obtained signal candidates, the detected signal is selected as the one that minimizes the log likelihood function. Furthermore, this paper proposes stream selection-MSCS (S-MSCS) that selects the constrained streams under the criterion of small equivalent amplitudes of channels caused by the MMSE detection. Setting the number of patterns of constrained streams to just one can reduce complexity, and selecting the constrained streams on the basis of the equivalent amplitude can maintain excellent BER performance. Computer simulations under 8 × 8 MIMO channel conditions with 16QAM demonstrate that S-MSCS can maintain only 0.5 dB degradation of the average BER performance from the maximum likelihood detection (MLD), while reducing the computational complexity to about one third of that of QR decomposition with M algorithm (QRM)-MLD.