Study On Signal Detection Algorithms Of MIMO System

Multi-input multi-output (MIMO) communication systems have already been proved have enormous spectral efficiency theoretically and experimentally. On the basis of MIMO, space-time coding combines channel coding and array processing, also greatly improving channel capacity and transmission rate of communication system. It is a suitable solution to the increasing demand for high quality wireless multimedia service. For a MIMO system, the signal on each receiver antenna includes the transmitted vector. These signals in time and frequency are overlapping, if the channel is frequency selective, there still exists inter-symbol interference. These all increase the difficulty of signal detection. Signal detection technology is to distinguish and restore the transmitted vector, and its performance will directly influence the whole system. Therefore, signal detection technology is one of the key problems of MIMO communication.There are many signal detection algorithms for MIMO system, Maximum likelihood (ML) detection is the optimal receiver, but its computational complexity increases exponentially with the number of transmit antennas and the constellation size. Many studies have been conducted to achieve low implementation complexity. Then many sub-optimal detection algorithms appeared, including: sphere decoding, linear detection algorithms, the VBLAST algorithm, algorithms based on matrix decomposition, Turbo detection, etc. For the VBLAST algorithm, all layers are detected by a successive interference cancellation technique which nulls the interferers by a nulling vector. Obviously an incorrect symbol selection in the early layers will create errors in the following layers, so the VBLAST detectors easily suffer from the error propagation. The performance of all system is worse by the error propagation. QR decomposition algorithm has attracted a lot of attention because they do not need to ask inverse matrix, the iterative QR decomposition algorithm has the problem of too many iteration times. We introduce the reliability criteria for its improvement. As the detection result meets the criteria, we stop the iteration. The improved algorithm reduces the iterative times, at the same time, keep the good performanceIn order to mitigating error propagation, we also present an improved group iterative QR decomposition algorithm. For the first QR decomposition, we just save the result temporarily for interference cancellation, but not making a decision. In order to enhance the performance, the iterative QR is applied. Then the layers with a good performance are kept, the other layers are grouped, the QR decomposition is applied again. The last detected layers get more diversity gain and less interferes, then the performance is proved. The improved algorithm achieves good performance to standard VBLAST with reduced computational complexity.