LABORATORY VERIFICATION OF MULTICHANNEL TRACKING OF DENSE GROUPS OF SMALL UAVS USING A DIGITAL TWIN OF A RADAR STATION

Abstract

The paper presents an experimental laboratory verification of the ability of a mobile airspace monitoring radar

to simultaneously detect and track dense groups of small multirotor unmanned aerial vehicles (UAVs). The study is performed using a previously developed digital twin of the radar, extended with a multichannel track processing layer. Four typical group-traffic scenarios are implemented in the twin: sequential entry of single targets, simultaneous and quasi-simultaneous appearance of several targets in the detection zone, and a combined group transit from three directions with up to 20 simultaneously tracked targets, inter-vehicle intervals of up to 5 s and inter-subgroup intervals of up to 10 s. The laboratory bench generates target trajectories with randomized speeds (8–42 m/s), altitudes (30–500 m), headings and manoeuvre patterns, after which track continuity, probability of track loss, the rate of false plot merging, and the latency of data delivery to external information consumers are estimated over fifty Monte Carlo runs per configuration. Track initiation uses an M-of-N confirmation rule, state estimation is based on an interacting multiple model filter, and two measurement-to-track association schemes are compared: global nearest neighbour assignment and joint probabilistic data association. Special attention is paid to plot-track association in dense formations, where mutual shadowing and unresolved plots provoke track swap and coalescence. The experiments show that the joint probabilistic scheme reduces the probability of track loss in the combined scenario from 0.117 to 0.064 and almost halves the false merging rate, at the cost of a 34 % increase of the processing time per scan. The computational load model verified against the bench indicates 94.8 % processor utilization at the design point of 20 targets and a 2 Hz update rate, which defines the required processing margin. The maximum divergence between the model and the bench does not exceed 9.6 % against the 15 % tolerance criterion, which confirms the adequacy of the digital twin. The engineering outcome is the confirmed multichannel capability of the system, justified tracker parameters and quantitative requirements for the processing hardware intended for airspace monitoring and the protection of critical infrastructure under dense group UAV traffic.

Keywords: unmanned aerial vehicle; digital twin; radar station; multichannel tracking; dense air traffic; measurement

association; track continuity; airspace monitoring.