Enhancement and assessment of a transient event characterization algorithm

Emerging technologies enable the acquisition of high rate temporal data over wide areas with moderate spatial resolution from remote sensors, either ground, air or space borne. These technologies are appropriate for collecting data on a wide range of brief events, including high explosive detonations, short-burn rocket motors, and muzzle flashes. The limiting factor in the interpretation of data acquired from such a system deployed to monitor battlefield activity is that the system may become overburdened by the number of events occurring in a rather short time period. Techniques to rapidly detect, identify and characterize these events are required prior to contemplation of such system deployment. An algorithm developed for performing Transient Event Characterization (TEC) from a passive remote sensor was presented in TEC99. This paper expands the concepts of event discrimination as well as identifying sources of uncertainty. Physical characteristics of an unknown event can be used to discriminate between event classes (e.g. artillery detonations, rocket boosters, and others), further discrimination can then be performed within each class. Regardless of class the event yield must be estimated, which requires some form of passive ranging in order to estimate the altitude of the event and thereby the line-of-sight atmospheric attenuation. This paper presents event class discrimination based on event waveform characteristics with applications to modeled data as well as a band optimization study for passive ranging.