Adaptive Doppler-shift Compensation for Ofdm Underwater Acoustic Communications System

An adaptive Doppler-shift compensation scheme for wideband orthogonal frequency division multiplexing (OFDM) is proposed in this paper. This approach accommodates a time-varying Doppler, where the time scaling factor for each OFDM symbol is frequently estimated depending on the inherent periodicity of the cyclic-prefix and its preceding neighbours' values. The proposed technique relies on a single preamble of a packet consisting of multiple OFDM symbols to detect the start of the packet, the concept of nearest neighbour rule to facilitate smoothing between symbols, and a dynamic symbol synchronization point to update the time scaling factor. To accomplish this, an adaptation step is derived involving the weight of the nearest neighbour's time scaling in estimating the current re-sampling factor. In this paper, the synchronization point is updated frequently and therefore an adaptive time scaling is obtained implicitly. The velocity perturbation is tackled by employing the concept of nearest neighbour rule in addition to adopting a dynamic synchronization point. Instead of utilizing the whole guard interval, the proposed system has exploited a finite length window of the cyclic prefix for correlation in each OFDM symbol to frequently estimate the Doppler shift. No iterative computation is required for the interpolation factor estimation. Furthermore, the proposed algorithm needs to only buffer one OFDM frame before data demodulation instead of buffering the whole data packet. Thus, it necessitates a lesser degree of complexity and memory requirements. An experiment has been conducted in the North Sea during 2009 and the proposed algorithm has been compared with the block Doppler compensation technique. Results show that there is variation in speed during the packet time, therefore the proposed system surpasses the block technique. It is confirmed that the time scaling factor of the adaptive system is estimated for each OFDM symbol, whereas the block approach fails in estimating these variations.