Real-Time Detection and Tracking of Multiple Ground Targets

Real-time detection of time-critical ground targets by wideband radar is a vital task for missiles and unmanned aerial vehicles. It is challenging to give consideration to efficiency and performance with low signal-to-noise ratio, non-uniform clutter distribution, radar cross section fluctuation, multiple-target overlapping, and interfering objects. This paper presents a complete method including integration, detection and tracking. First, an improved non-coherent integration algorithm based on velocity prediction via Keystone transform is designed to achieve fast energy focusing of weak targets. Then, an anti-false-alarm CCA-CFAR based on clutter energy balance process is used to extract reliable observations. Finally, a real-time multi-target tracking and error correction algorithm based on the estimation of the number of targets is proposed, which realizes real-time target tracking and elimination of false alarms, and reduces the missed detection caused by the phenomenon of radar cross section fluctuation and multiple-target overlapping. Experimental results based on practical data show that the proposed method effectively reduces false alarms and missed detections, and achieves real-time, continuous and accurate detections under complex condition.