Differential orthogonal frequency division multiplexing (OFDM) is practicallyattractive for underwater acoustic communications since it has the potential toobviate channel estimation. However, similar to coherent OFDM, it may sufferfrom severe inter-carrier interference over time-varying channels. To alleviatethe induced performance degradation, we adopt the newly-emerging partial FFTdemodulation technique in this paper and propose an eigendecomposition-basedalgorithm to compute the combining weights. Compared to existing adaptivemethods, the new algorithm can avoid error propagation and eliminate the needfor parameter tuning. Moreover, it guarantees global optimality under thenarrowband Doppler assumption, with the optimal weight vector of partial FFTdemodulation achieved by the eigenvector associated with the smallesteigenvalue of the pilot detection error matrix. Finally, the algorithm can alsobe extended straightforwardly to perform subband-wise computation to counteractwideband Doppler effects.
展开▼
