Layered Structure ML Detection Scheme for MIMO Systems

Recently, Multi-Input Multi-Output (MIMO) systems attract attention and the detection schemes for MIMO systems are investigated actively. Among these schemes, Maximum Likelihood (ML) detection scheme can provide optimal performance, however its complexity is so enormous that it is difficult to implement it. On the other hand, Ordered Successive Interference Cancellation (OSIC) scheme which is a detection part of Bell-labs LAyered Space-Time (BLAST) system reduces the burden of complexity at the cost of performance. In this paper, to reduce the computational complexity without significant deterioration of performance, we propose a Layered Structure ML (LSML) detection scheme which is the hybrid scheme of ML and OSIC scheme. The proposed scheme does not detect a sub-stream symbol at each stage like OSIC, but choose several surviving symbol vectors at each stage, fed them to the next stage, and finally detects all sub-stream at last stage. According to simulation results, the error performance of LSML converges quickly to that of ML, as the number of surviving symbol vectors increases. And the computational complexity of LSML is 0.043 to 0.12 times that of ML, depending on the number of surviving symbol vectors, when the numbers of transmit/receive antennas are 4 respectively and QPSK-modulation is used.