Optimal Resampling of OFDM Signals for Multiscale–Multilag Underwater Acoustic Channels

Abstract

Underwater acoustic channels (UACs) have a large fractional bandwidth and are well modeled by a multiscale–multilag model, which is a generalization of the commonly employed single time-scale model. The path-dependent time scaling introduces significant channel variation, destroying carrier orthogonality and yielding intercarrier interference (ICI) in an orthogonal frequency division multiplexing (OFDM) system. Typical front-end processing at the receiver for a channel with common time scaling on every path involves resampling the received signal to compensate for this scaling. This work examines the choice of resampling parameter for the multiscale scenario. Bounds are derived for the performance and hard and soft data detection schemes using the Hammersley–Chapman–Robbins bound (HCRB) and the Cramér–Rao bound (CRB), respectively, which when optimized yield the optimal resampling parameter. It is shown that resampling is equivalent to matched filtering the incoming signal with an approximated channel having a common time scale on each path. A second criterion for optimizing the resampling parameter then minimizes the aggregate error between the multiscale channel and its single-scale approximation. Analysis of special cases suggests that the optimal resampling parameter is close to the time-scaling parameter of the dominant path. The scenarios under which using the time-scaling factor on the dominant path for resampling are suboptimal are also investigated. Blind and pilot-aided estimators for the optimal resampling parameter, which are a generalization of the estimators for the common time-scale model, are derived building on the channel approximation model for resampling. Numerical simulations show the advantages of the derived estimators over a classical packet-length-based time-scale estimator.