Transient signal characterization using multi-lag phase space analysis

Transient signals have proved to be difficult to characterize/classify because of their short duration and broad spectrum. Methods based on time-frequency analysis are used for transient detection, but their characterization is limited when it comes to similar signals that have close amplitudes and frequencies. This paper presents the multi-lag phase space analysis, an alternative to the classical approaches for the characterization of transient signals. New tools developed for this method are applied on multi-path acoustic signals which are acquired using an acoustic microphone placed in the center of a facility surrounded by walls. The results are compared with the ones obtained using time-frequency analysis.