A non-data-aided algorithm based on ML for OFDM synchronization

Orthogonal frequency division multiplexing (OFDM) is more sensitive to symbol timing offset (STO) and carrier frequency offset (CFO) than single-carrier signals, which requires more accurate synchronization algorithms. In this paper, the traditional ML synchronization algorithm is improved by accumulating multiple OFDM symbols, and joint estimation of symbol timing offset and carrier frequency offset is accomplished without data aiding. The simulation illustrates that the improved ML algorithm has a higher accuracy for STO estimation and a lower MSE for CFO estimation.