Research On Signal Detection Algorithms For Distributed MIMO System

A new communication structure which combines the traditional point-to-point MIMOtechnology with distributed antennas technology is intituled distributed MIMO system.Compared with the traditional centralized MIMO system, distributed MIMO system has theadvantages of lo transmit power, lagre system capacity, wide coverage, and low radiation on thehuman body, which has been proposed as one of the candidate systems in the future wirelessmobile communication systems.Signal detection is one of the most important techniques in the receiver of distributedMIMO system. Because of the antennas distributed in dieffrent locations, the detectionalgorithms based on the channel model of centralized MIMO can not be direct applied indistributed MIMO system. In order to solve the problem, this thesis firstly gives the signaltransmission model of distributed MIMO system, and then based on this point, studys on signaldetection algorithms of distributed MIMO signal. The main contents are as follows:1. According to the wireless channel transmission characteristics, analysising thecharacteristics and differences of centralized and distributed MIMO systems and discussing thedistributed MIMO signal transmission model for the foundation of researching detectionalgorithms.2. Aiming at the high complexity problem of ML detection algorithm, a simplifiedalgorithm based on traversal searching part symbols is presented. This algorithm detects partsymbols by the whole space traversal searching, each searching uses low complexity algorithmto detect the remaining symbols, and calculates the Euclidean distance between estimatedsymbol vector weighted with the received symbol vector. After that, finding the minimumEuclidean distance corresponding symbol vector is transmitted symbol vector estimation.Theoretical analysis and some computer simulation results show that this algorithm can achievenear the ML algorithm detection performance with low complexity.3. When the channel transport conditions deteriorated, by introducing the idea of Turboiterative decoding to improve system detection performance. Aiming at the defect of thetraditional MMSE iterative detection algorithm which has slowly convergence rate and needsmany iterations to get better BER performance, a new soft detection algorithm based on posteriorinformation iterating is proposed. This algorithm can obtain better detection performance withless iterations by using the previous iteration posterior information provided by soft channeldecoder feedbacked and current iterative estimated symbol external information provided by softdetector to calculate the remaining symbols external information. Some computer simulationresults show that the algorithm can expedite convergence rate of iterative detector and reduce thesystem complexity remarkably.4. In some cases, it is difficult to accurately estimate the channel or accurate estimate thechannel with the high cost. In these oeeasions, differential detection technology without channelinformation is an attractive seheme. Aiming at the limitation that the channel fading factor is required to obey the quasi static distribution of traditional scheme can correctly detecte symbols,a differential detection scheme is designed. In order to reduce the fast change channel impact tosystem BER performance, this scheme restructures the sending space time symbol matrixs andmakes the symbols which has the differential phase relation be sent out in the before and afterthe two times. Simulation results in fast fading channel indicate that this scheme can improve theBER performance remarkably compared with the conventional scheme.