Spectral Modelling and Low-complexity Blind Carrier Frequency Tracking in OFDM

Satisfactory operation of orthogonal frequency-division multiplexing (OFDM) requires carrier frequency estimation with precision on the order of 1 to 2% of the subcarrier spacing. Numerous pilot-aided and blind algorithms that achieve this precision have been proposed. Blind algorithms are generally more complex than the pilot-aided algorithms. In this paper, we propose a novel low-complexity blind algorithm that takes advantage of a key property of the in-band OFDM power spectral density (PSD). To date, OFDM PSD research and modelling has mainly been directed towards reducing the out-of-band radiation. Here, we turn our attention to the in-band portion of the spectrum and show that it has an oscillatory variation. We show how this property can be used for carrier frequency tracking. Computer simulations show that the proposed algorithm works well in high-noise, frequency-selective channels.