Skip to Main Content
This paper describes a statistical signal-processing method for exploiting narrowband bistatic RF measurements to detect and track moving people (hereafter referred to as “dismounts”). In our approach, RF measurements are made by a constellation of narrowband radar units, arranged around a surveillance region. There are several benefits of narrowband radar in this application, which we describe in the paper. However, the narrow bandwidth means that individual measurements only yield coarse information about target state. We show that by fusing measurements from multiple bistatic sensors over time with a Bayesian nonlinear-filtering algorithm, we can effectively estimate dismount position and velocity using as little as 5-10 m bistatic range resolution. We illustrate the algorithm's efficacy with an experiment where a moving person is detected and tracked from a constellation of four narrowband bistatic sensors.