Research On Channel Estimation Technique And Decoding Algorithm In MIMO-OFDM System

Along with the society incessant requirement, the transmission speed isenhanced incessantly in communication, and the scarcity of frequency resource isapparent to all of us day by day, thus some disadvantage in transmission systemhave emerged, especial in seriously inter code interference fading channel, seriallytransmission in high speed becomes very difficulty. The data stream is divided intoseveral bit streams in OFDM system, so every subdata stream is transmitted inmuch slower speed. MIMO technology has achieved transmission division andreceiver division basing on space-time code and multi-elements antenna. Hence wecombine OFDM with MIMO, which resists inter code interference and multi-pathsfading more effectively.Exact channel estimation is the key to guaranteeing the data streams to betransmitted safe and the MIMO-OFDM in good performance. The paper analyzesthe channel estimation scheme based on the scatter pilots, and the computersimulations demonstrate the effectiveness of the scheme. The theoretical analysisand numerical results show that the scheme transfers multi-antennas channelestimation into single antenna channel estimation, which can be achieved simplyand effectively, besides these, the scheme is of stable estimation performance andstrong faculty in following up the variational channel. However, because of lowbandwidth efficiency, the scheme is not applicable in many antennas condition.Aiming at the two great challenges: bandwidth efficiency and multi-pathsfading, we propose a blind channel estimation algorithm. By combining subspacemethod with the special structure of space-time block code and the speciality ofOFDM, we accomplish the blind channel estimation, which does not require anytraining symbol. Finally, combining the computer simulations, we compare thealgorithm with the quasi-channel and channel estimation by pilots in performance,and we conclude that the channel estimation by pilots is close to the quasi-channel,but the pilot symbols expend the quarter of the bandwidth resource. The blindchannel estimation does not use training symbols, there is loss in performance, butthe error because of the channel estimation affects the decoding performance verysmall, which heightens the transmission speed.Obviously, our object to channel estimation is decoding correctly, which is thatwe obtain the input sequence of the transmitting antennas making use of thechannel state information and the output sequence received. About decodingmethods, the paper analyses and discusses the maximum likelihood decodingalgorithm, and educes the decoding algorithm for the MIMO-OFDM coding matrixconstructed in the paper. In practice, the channel is frequency-selective fadingchannel, OFDM divides the channel into several flat fading subchannels, thenumber of the subchannels is large relatively, which increases the decodingoperation linearly. To low the operation and the affection to decoding by the errorof channel estimation, we propose a new decoding algorithm, that is pseudoinversealgorithm. Pseudoinverse decoding obtains the optimal result in Least Square sensebased on the transmitted symbols. In the fast-stepping algorithm, we make use ofthe iterative grads method, which weakens the affection of the channel estimationerror to the decoding performance.Broad band multimedia communication is the inevitable direction of the futurecommunication, frequency's reasonable use and the data transmission speed aretaken into our account firstly. Aiming at the two challenge, the paper propose ablind decoding algorithm, which does not require training sequence and channelestimation. In despite of some loss in performance, the algorithm increase thefrequency spectrum's using efficiency effectively, and the operation is alsopredigested much. It is apparent that the blind decoding adapts to the developmentof the future mobile communication, how to optimize the algorithm's performanceulteriorly is the research direction for the future absolutely.