Signal analysis approach for passive tomography : applications for dispersive channels and moving configuration

Underwater channel is an example of a natural environment potentially characterized by signals generated by various sources : underwater mammals, human activity noise, etc. In order to take advantage of these sources, the concept of passive acoustic tomography has been introduced. According to this concept, the environment parameters could be extracted from the analysis of the received signals. While the signal’s parameters are intimately related to physical parameters of the environment, their accurate extraction is crucial. That is, this task is complex while we work in completely passive context and when we deal with a large diversity of underwater signals. Generally, signals issued from underwater environment have complex time-frequency structures : non-linear timefrequency and multi-components. Two typical non-linearities are generated by the dispersive systems and the relative motion between transmitter and receiver. Despite the origin of these phenomena, the signal approach proposed in this paper will provide a unique framework for parameters extraction. This approach is based on the time-frequency-phase coherence of any natural non-linear time-frequency component. Taking advantage of this coherence, the nonlinear structures can be efficiently extracted and used for physical parameters estimation. .