The University of British Columbia

The paper presents a low complexity discrete stochastic approximation algorithm for time and frequency synchronization in OFDM systems. The proposed technique can track the conditions of a slowly time varying channel where synchronization parameters, namely the symbol timing and frequency offset, vary slowly with time. The most important property of the proposed algorithm is its self-learning capability - it spends most of the computational effort at the global minimizer of the objective function. In particular, we show that the algorithm achieves an #949;=(1- rho;)/(1+ rho;) reduction in computational cost (in terms of complex multiplications to be performed), where rho;=Ncp/N is the ratio between cyclic prefix length (Ncp) and number of subcarriers (N) of the OFDM system (e.g. N=512 and Ncp=64 gives #949;=78%). Numerical examples illustrate the synchronization accuracy of the proposed technique in terms of symbol timing and frequency offset estimation errors.