Research On Channel Estimation Techniques In OFDM Communication System

Due to the high spectrum efficiency, ease of implementation (via FFT), and low receiver complexity (single-tape equalization), etc., orthogonal frequency multiplexing technique (OFDM) has been adopted and is currently considered by many application standards, including digital audio and video broadcasting (DAB, DVB) in Europe, high-speed ADSL modems over twisted pairs in the United States, local area mobile wireless broad-band standards IEEE802.11a, MMAC, HIPERLAN/2, and the fourth generation mobile system.Channel estimation plays a very important role in OFDM system. By channel estimation, coherent detection can be carried out which gives better performance in low SNR. This thesis investigates a variety of issues involving channel estimation in OFDM system:First, we discuss how to design the parameter of OFDM system in this paper and give the theoretical result for OFDM system design. Those discuss result are verified by simulation in this paper. We show the relationship between inter-symbol interference and protect prefix. The length of protect prefix need not to be longer than the channel multipath length. As long as the channel power inside the protect prefix take more than 97% of the total channel power. The pilot pattern is also considered in this chapter. The optimum pilot pattern has the scattered type. The pilot space should base on the fourth-order moments of the Doppler spectrum.Second, in this thesis we discuss how to make OFDM system combats with time selective channel. We propose a channel estimation method with hidden pilot in this paper. Because no separate pilot is used for channel estimation, channel band is saved. As the simulation shows, hidden pilot can save 9% 11% of the channel band. Hidden pilot is a kind of coherent detection. Different from those channel estimation method base on blind detection, hidden pilot use pilot to carry out channel estimation. It has a faster convergence rate. After comparison we find hidden pilot has a similar performance with MMSE channel estimation method. Cramer-Rao low bound of this method is deduced in this paper. Third, we discuss the equalization technique for OFDM system in time-variant mul-tipath chanels. Due to the expansion of symbol length, OFDM system are very sensitive to time variation channel. Time variations corrupt the orthogonality of the OFDM subcar-rier waveforms so that intercarrier interference (ICI) occurs. We propose a low complexity equalizer-LCD-FFT equalizer. As shown in this paper, channel will vary with linear fashion in most circumstance. Compare with other equalizer base on channel shorten technique, LCD-FFT equalizer is more accurate. Because the LCD-FFT equalizer is based on FFT, it has o(Nlog₂N) complexity. Simulation result shows that LCD-FFT equalizer has an similar performance with LS equalizer which has o(N³) complexity.The last one is how to carry out channel estimation in MIMO-OFDM system. Multi-antenna systems are the spread of single-antenna systems. When the transmitter and the receiver use more than one antenna, MIMO channel is constructed. As information theory indicates, MIMO channel has larger capacity. This means multiple-antenna systems can transmit information with higher data rate when the bandwidth and signal-noise-ratio is fixed. In this thesis, we propose a simplified MIMO-OFDM channel estimation method base on MMSE criteria.