Time-frequency-phase coherence - general framework for signal analysis in passive context

Abstract

The characterization of a natural environment (underwater, for example) and the identification of radar/communication signals in SIGINT (signal intelligence) are just two typical examples of applications requiring signal analysis in a passive configuration. In the first case, even if the characterization is based on the analysis of received signals in an active configuration, the unknown deformations of the transmitted signal transform the signal processing problem in a passive context one. Concerning the second case, the passive behavior of the signal intelligence field is a well-known problem in the electronic warfare problem.In this paper we propose a general signal analysis framework in passive context. We show that, in spite of the differences between some possible passive applications (underwater channel characterization and SIGINT) a unified signal analysis framework can defined. This definition starts from the general observation that real life signals received in a passive configuration are non-stationary. Their analysis in the time-frequency domain is well adapted so that it offers appropriated structures which are good candidates for the information post-processing. In a passive context, the definition of an appropriate time-frequency representation space is a complex problem, mainly related to the lack of a priori information about the processed signal. One general solution is proposed in this paper and it is based on the time-frequency-phase coherence. Conceptually, while the received signals are unknown (a model is difficult to be assumed), a general remark is the coherent shapes of their time-frequency structures. This coherence could be materialized by fundamental physical parameter of every signal - amplitude, time, frequency and initial phase. Indeed, the signal analysis framework is defined through three blocks : detection of regions of interest, segmentation and separation, analytical characterization. This architecture is mainly based on joint use of time, frequency and local phase analysis. More precisely, the phase information will be locally analysed, using generalized instantaneous moments, on the time-frequency regions previously selected thanks to the time-frequency grouping algorithm. This architecture constitutes an efficient scheme to solve the constraints brought by this type of signals with a complex time-frequency behavior and by the human operator to reduce his tasks in the decision process. Examples from underwater behavior (underwater mammals vocalizations) and electronic warfare will prove the efficiency of the proposed approach.