Superposition Of Training Sequences For Mimo-ofdm Signal Design And Testing Of Key Technologies

Multiple-input multiple-output (MIMO) can be used in orthogonal frequency division multiplexing (OFDM) systems to not only offer the potential of high spectral efficiency and the promising to exploit maximum spatial resources, but also lessen the severe effects of frequency-selective fading. Therefore, MIMO-OFDM is the one of key techniques of the next generation mobile communications system.First of all, this paper surveys the research status of the channel estimation of MIMO-OFDM system. The disadvantages of time divided, frequency divided, and code divided pilot based channel estimation scheme are discussed, and the problem is proposed.In the second section, a channel estimation scheme based on the superimposed training sequences is proposed for OFDM system. The closed form expression of bit error rate (BER) is presented when using the first order statistics of superimposed sequences to estimate the channel response vector. And the simulations validates the analysis results. Compared to the pilot based scheme, the main differences of the proposed scheme is that the superimposed training sequence is completely overlapped with the data in the time domain, the frequency domain, and the code domain. Therefore, the data transmit rate will not be reduced due to pilots in the proposed scheme, and time and frequency synchronization can be achieved in one OFDM symbol.According to analysis and simulations, the BER performance of OFDM system with superimposed training sequence is affected by the training sequence separation factor and power allocation factor. Based on the minimum error rate principle, the closed form expressions of the optimal training sequence separation factor and the optimal sequence power allocation factor are presented in the third section. The analysis and simulations show that: 1. The optimal separation factor is determined by the signal to noise ratio (SNR) and the power of superimposed training sequence, and when SNR is go higher, it should be reduced; 2. The optimal power allocation factor is determined by SNR, and when SNR is go higher, it should be increased.The channel estimator is severely interfered by data, when using the first order statistics of the training sequence to obtain the channel state information. And two iterative schemes are proposed based on the ideas that reconstructing the transmitted signal and reducing the interference of data in the received signal, respectively. According to simulations, compared to the ideal channel estimation, the proposed schemes can approach to the performance bound, and when BER is 10?2 , the performance loss of the first scheme is about 1dB, and that of the second scheme is 2dB.Finally, this paper focuses on the performance analysis of distributed transmit antennas system (DTAS). The discussed system is affected by different random propagation delay and log-normal shadowing in each spatial multiplexed data stream, and zero-forcing detection is adopted to recover the data. The BER performance of DTAS and co-located transmit antennas system (CTAS) are compared by analysis and simulations in composite shadowing and Rayleigh fading channel. It is shown by analysis and simulations that DTAS can perform better than CTAS when SNR is higher than a threshold, and the threshold is determined by space-time block length, propagation delay, and standard deviation of shadowing.The proposed MIMO-OFDM system with superimposed training sequence can provide much flexible and efficient access for the next generation wireless communication system. The superimposed sequence can simultaneously achieve the time and frequency synchronization and channel estimation. Moreover, the proposed scheme is very easy to promote to fast fading channel, if the length of the sequence is shorter and more than one sequence can be superimposed in one OFDM symbol.