Secure positioning in a UAV swarm using on-board stereo cameras

The widespread use of quadcopters as unmanned aerial vehicles (UAVs) provides a number of application possibilities, mostly using a collaborative swarm of small UAVs to optimise critical missions. The common application domains for small UAV swarms are surveillance, path planning, air-bone and as relay networks. Cooperative applications make use of the UAVs' locations to make decisions. However, security vulnerabilities must be considered when the system infers rights based on the UAV's location. An attacker can cheat the system by declaring a false or inaccurate location to gain access to resources that are restricted or to undertake malicious activities without detection. In this paper, we propose the use of a UAV payload stereo camera to measure the distance between a UAV, called a prover node, and the set of closest verifier nodes. Using a multilateration algorithm to estimate the prover's position, we performed simulations for two cameras with different capabilities. The simulations showed that the proposed technique provides 98% validation accuracy for distances up to 50 m for 1280x720-resolution cameras and more than 99% validation accuracy for distances up to 100 m for 1920x1080-resolution cameras. For distances greater than 100 m between the prover and the verifiers, the accuracy depends exclusively on the stereo camera with the highest capacity.