The Impact of Initial Swarm Formation for Tracking of a High Capability Malicious UAV

The use of UAVs or drones for criminal or terrorist enterprises is an increasing problem. Many countermeasures have been proposed to prevent, deter, detect and/or mitigate the dangers posed by such malicious UAVs. One such countermeasure is to track or pursue a malicious UAV back to its point of origin using one or more surveillance UAVs in order to apprehend the UAV and possibly its owner. If the malicious UAV has a higher capability set than the surveillance UAVs, it will be able to outrun any one of them, and therefore the tracking responsibility must be distributed over a swarm of surveillance UAVs that are geographically dispersed across the tracking area of interest. One aspect of particular interest is how the initial formation of the swarm of surveillance UAVs impacts its ability to successfully track a malicious UAV. In this paper, we examine a specific circular initial swarm formation comprising uniformly spaced concentric rings of uniformly spaced UAVs. The total number of surveillance UAVs follows the sequence of centred hexagonal numbers as the number of rings increases. The tracking performance of this circular swarm of surveillance UAVs is compared to a reference swarm of the same size in which the initial locations of the UAVs are randomly chosen. Two tracking strategies are considered: 1) Reactive tracking, in which each surveillance UAV acts independently of the others and only pursues the malicious UAV when it itself detects it, and 2) Reactive tracking with predictive pre-positioning, in which once one surveillance UAV detects the malicious UAV, it communicates the estimated trajectory and speed of the malicious UAV to all swarm members so they can predictively move to a more optimum tracking position before the malicious UAV arrives. The results demonstrate that this particular circular swarm of surveillance UAVs has superior tracking performance relative to the reference randomly positioned swarm of the same size; this is true for both tracking strategies, but particularly when predictive prepositioning is employed with a relatively small number of surveillance UAVs.