Research On Joint Channel Estimation And Signal Detection For MIMO-OFDM System

The performance and complexity of receiver signal detection algorithm has a significant impact on the reliability and real-time performance of MIMO-OFDM system.In general,signal detection algorithm that owns a good detection performance has a higher complexity,so it does not apply to the actual.Therefore,it is very important for MIMO-OFDM receiving system design to find a kind of signal detection algorithm which owns a good detection performance and lower complexity.Moreover,only channel is accurately estimated that receiving terminal could have effectively removed the interference caused by channel.So channel estimation is also one of the most important technologies in MIMO-OFDM system.Channel estimation and signal detection are usually operated independently in MIMO-OFDM systems.The precision of channel state information depends on the training sequence.The system BER performance will be decreased if the precision is low.In addition,channel coding is not considered in most signal detection algorithms and the hard-decision is applied in these detection algorithms.Meanwhile,some useful information from decoder is not used in these algorithms.Therefore,the system performance can be improved effectively when channel estimation,signal detection and decoding are combined by iteration.This paper has deeply researched joint channel estimation,signal detection and channel decoding in the MIMO-OFDM system.The main research result are as follow:Firstly,we have researched the structure,the optimal pilot design method and channel estimation algorithm with auxiliary pilot of the MIMO-OFDM system as well as giving the corresponding theoretical analysis.Moreover,because blind channel estimation algorithm has a slow rate of convergence and singular value decomposition brings in high complexity in the MIMO-OFDM system,a new half-blind channel estimation algorithm is proposed,which is based on the modified subspace of MIMO-OFDM.This kind of algorithm traces the noise subspace to estimate the channel state information with fast data projection method instead of singular value decomposition.Then,using training sequence checks channel to recognize fuzzy matrix that is hardly ensured with blind channel estimation algorithm,which improve the precision of channel estimation algorithm.Simulation result has found out the new algorithm not only has a faster convergence speed,but also decreases algorithm complexity.So the half-blind channel estimation algorithm can be adopted.Secondly,the layered space-time coding,classical MIMO systems' signal detection technology and its corresponding theoretical analysis and simulation are discussed.On the basis,a quantum search algorithm will be introduced into the detection system and realizes a improved signal detection quantum search algorithm with low complexity theory.Simulation results show that the algorithm can approximate the detection performance of the maximum likelihood detection algorithm,and the complexity of the system has been reduced significantly.To solve the problem of selecting initial radius of the sphere decoding algorithm and the final output judgment,an improved sphere decoding detection algorithm that suitable for high signal-to-noise ratio is proposed.For the fact that the Particle Swarm Optimization signal detection algorithm is easy to fall into local optima and has a low convergence accuracy,an improved particle swarm optimization signal detection algorithm that suitable for low signal-to-noise ratio is proposed.Combined with the hybridization technique of the Genetic Algorithms and extreme perturbation mechanism,the new algorithm has been improved towards the old one.Simulation results show that the algorithm can effectively improve the signal detection performance of the system and has a strong global searching capability.So it is a practical method of signal detection.Once again,theoretical researches on expectation maximization algorithm and its extension are further stretched.On the basis of the researches,channel estimation methods in time and frequency domain are proposed based on the maximized iterative expectation of generalized space.Combined with signal detection theory,a frequency domain SAGE-ISD joint channel detection and estimation algorithm is presented concerning the relatively slow convergence rate adopting expectation-maximization algorithm.The proposal primarily adopts linear minimum mean-square error channel estimation to evaluate the channel state.Then by means of joint iterative technique combined with generalized space iterative expectation maximization algorithm and improved sphere decoding algorithm,realize channel estimation and signal detection,thus improve system stability.Since maximum likelihood detection algorithm in time domain is rather difficult for hardware implementation and the traditional channel estimation is of relatively poor performance,a joint channel estimation methodology based on time domain SAGE-IPSO is proposed.The program utilizes a set of training sequence to achieve preliminary channel evaluation.Adopting generalized space iterative correction expectation maximization algorithm finished channel estimation calibration,further more,through improved particle swarm optimization algorithm,complete iterative detection of the signal,therefore enhance system performance.The simulation results demonstrate the above two proposals can achieve channel state information and detection data estimation with fewer iterative calculation,in addition,largely approximate the maximum likelihood detection algorithm under ideal channel estimation.Finally,we have researched the channel coding and decoding algorithms as well as soft input soft output signal detection.Besides,an improved iterative joint receiving scheme is proposed for the MIMO-OFDM system,which is used to solve higher error rate in traditional MIMO-OFDM receiver where channel decoding,channel estimation and signal detection work independently.The new scheme can joint conduct channel estimation and signal detection using the soft information provided by the decoder.Simulation result has proved the proposed scheme prominent increases the performance of the system.