Sensor Adaptive Target Tracking over Variable Bandwidth Networks

Abstract

Tracking a maneuvering target is often formulated as a problem of estimating the state of a partially observed jump Markov linear system. For example, consider a radar track-while-scan system where a stream of range and bearing (and perhaps doppler) measurements are obtained at regular intervals (the radar scan rate) on all targets (aircraft, clouds) within the field of view. The purpose of the tracking system is to search through this data in real time for trails of measurements which correspond to targets of interest (aircraft) and then to follow these targets and estimate their kinematic parameters (range, speed, heading, etc). False tracks can arise from consistent trails of noise and cl..