Clustered Adaptation for Estimation of Time-Varying Underwater Acoustic Channels

In this paper, we present a sparse channel estimator that exploits channel coherence across blocks for orthogonal frequency division multiplexing (OFDM) in underwater acoustic (UWA) transmissions. We propose a novel channel variation model that the channel paths within a cluster share the same amplitude, delay, and Doppler scale variations from one block to the next, but different clusters vary independently. The variation offsets for different clusters are estimated based on the measurements on the pilot subcarriers of the current block. A hybrid channel estimator is developed that can effectively utilize both the channel knowledge from the previous block and the pilot observations of the current block. Performance results based on collected data from two experiments demonstrate that adaptation with multiple clusters outperforms that with one cluster, and that the proposed cluster-adaptation based channel estimator significantly improves the system performance relative to the pilot-based counterpart.