Using Phase Space Methods for Target Identification

Abstract

The response of a radar or sonar target to a signal may be described by an impulse response function, which means that the target may be considered as a filter acting on a signal. It is known that filters are not exactly invertible, and this lack of invertibility may be used to identify the particular target that reflected a signal. We apply techniques from nonlinear dynamics to determine the probability that a function exists between 2 signals. If 2 identical signals are filtered by the same filter, then our statistic will indicate a high probability that a function exists between the 2 signals; if the 2 signals were filtered by different filters, then the statistic will show a low probability that the 2 signals are related by a function. We demonstrate target identification with both numerical simulations and acoustic experiments.