Comments: download file (pdf)
Ref.: PhD Thesis, University of Algarve, August 2017
Advances in underwater acoustic communications require the development of methods to accurately compensate channels that are prone to severe double spreading of time-varying multipath propagation, fading and signal phase variations. Assuming the environmental information as a key issue, this work aims to improve communications performance of single-input-multiple-output links in such channels through the enhancement of their estimates used for equalization. The acoustic propagation physical parameters of the environment between the source and the receivers are considered in the process. The approach is to mitigate noise effects in channel identification for Passive Time-Reversal (PTR), which is a low complexity probe-based refocusing technique to reduce time spreading and inter-symbol interference. The method Environmental-based PTR (EPTR) is proposed that, inspired by matched field inversion, inserts physics of acoustic propagation in the channel compensation procedure through ray trace modeling and environmental focalization processing. The focalization is the process of tweaking the environmental parameters to obtain a noise-free numerical model generated channel response that best matches the observed data. The EPTR performance is tested and compared to the pulse-compressed PTR and to the regularized `1-norm PTR. The former is based on classical `2-norm chan- nel estimation and the latter, inspired by compressive sensing, uses weighted `1-norm into the `2-norm estimation problem to obtain improved estimates of sparse channels. Exper- imental results succeed with the proposed method for signals containing image messages transmitted at 4 kbits/sec from a source to a 16-hydrophones vertical array at 890 m range during the UAN'11 experiment conducted off the coast of Trondheim (Norway). The scientific contributions of this work are (i) the understanding obtained on the process of employing physical modeling and environmental focalization to equalize and retrieve received messages in underwater acoustic communications, thus exploiting the sensitivity of environmental parameters in order to adapt a communications system to the scenario where it is used; and (ii) the presentation of a new PTR-based method that focuses environmental parameters to model suitable noise-free channel responses for equalization and whose real data results succeeded for a set of coherent signals collected in experiment at sea. The proposed method contributes in a step forward to better understand how to insert physical knowledge of the environment for equalization in digital underwater acoustic communications.
Keywords: Underwater acoustic communications, passive time reversal, equalization, coherent communications, physical modelling, environmental focalization.