Research On Joint Synchronization And Channel Estimation Algorithm For OFDM Systems

Orthogonal frequency division multiplex (OFDM) is a multicarrier modulationtechnology with high data transmission rate. OFDM has high frequency efficiency,and is robust to multipath fading and narrow-band interference. According to theseadvantages, OFDM has become to the one of the core technology in the nextgeneration wireless communication.When transmitting in wireless communication channel, OFDM signal will sufferfrom severe environment included power attenuation, multipath fading and timingvarying. At the receiver, the existed timing errors and frequency offsets in receivedsignals will cause harmful effect. Meanwhile, the signals have distortion when afterwireless channel transmission. Thus the channel parameters should be estimated forchannel capacity improvement and correct coherent demodulation execution in OFDMsystems.In this paper, the synchronization and channel estimation algorithms areinvestigated in OFDM systems, and the main works are shown as follows:(1) The traditional symbol timing synchronization algorithms are studied.Schmidl’s method estimated the timing offsets by utilizing two repetitive trainingsymbols. Its timing metric had a peak plateau, which would lead to the uncertainty oftiming point and just suited for coarse synchronization. Park and Ren improved theschmidl’s method respectively, and could get a sharper timing metric. However, theywere not robust to Rayleigh multipath fading channel. The Hamed’s method workedwell in Rayleigh multipath channel, but its computing complexity is much higher.This paper proposes a new symbol timing synchronization algorithm by well using theproperty of the Golay complementary pair. The new method can still obtainimpulsive-like timing metric in multipath fading channel, which can greatly improvethe timing accuracy with low computing complexity and suits for coarsesynchronization and fine synchronization.(2) Least square (LS), discrete fourier transform (DFT) and linear minimummean square error (LMMSE) channel estimation algorithms are discussed in this paper.LS method was the easiest way for channel estimation. But without considering noiseinterference, the estimation accuracy was unacceptable. DFT and LMMSE methodimproved the LS channel estimates by reducing the noise samples. LMMSE offered the best mean square error (MSE) performance. However, it required the channelpriori knowledge and a heavy computational load. Most of the channel estimationmethods assume that there are no synchronization errors. In this paper, a joint timingsynchronization and channel estimation algorithm is proposed. The initial channelimpulsive response samples obtained in timing synchronization procedure areexploited to detect the first channel tap and adjust the timing point. These works arefinished in time domain, after remove the noise samples, the channel frequencyestimates in frequency domain are acquired by DFT. The new joint algorithm can getwell MSE performance with low system overhead and computing complexity.