Exploiting the Momentary Dependence of Radar Observations for Non-Cooperative Target Recognition

Multiple radar sensors can be used in collaboration to detect targets in an area of surveillance. In this paper, we consider a case, in which a target is detected by a network of radars producing multiple observations of the radar signature of the target during a short time window. Given that this time window is sufficiently narrow, the observations have a dependence between them momentarily related to the change in the orientation of the target. We propose the fusion of these interdependent observations to aid target identification by forming a joint multi-dimensional histogram of the radar cross section (RCS). In addition, we investigate the criteria for windowing the observations to ensure adequate interdependence. We present a case study to demonstrate the ability of the proposed approach to distinguish between different targets using the measured RCS collected by a multi-radar surveillance system. Based on the experiment, we analyze the criteria for the dynamic windowing and discuss the computational requirements of the proposed concept.