A Novel OFDM Symbol Synchronization Algorithm Based on Cyclic Prefix in DVB-T

Van de Beek presented the joint maximum likelihood estimator of time and frequency offset in OFDM systems. In most situations, its´ performance suffices in an acquisition mode. But in low signal-to-noise ratio AWGN channels and multi-path fading channels, the maximum likelihood peak values are not stable in the right place due to high noise effects, multipath delays and Doppler effects respectively. Some samples of cyclic prefix have been polluted by ISI. A new improved solution based on cyclic prefix in DVB-T system is proposed in this paper, which chooses different methods according to signal-to-noise ratio and multi-path delay depth. Averaging the maximum likelihood function in several consecutive symbols, dynamic regulation correlation slide windows and enlarging observation vectors are the key optimized points. It is shown that the proposed method can get sharp peak values and improve the symbol precision compared with the existing methods both in AWGN channels and in multipath channels.