An Efficient Path Planning Algorithm for 2D Ground Area Coverage Using Multi-UAV

Abstract

Unmanned aerial vehicle (UAV) equipped with visual sensors are extensively used in area coverage applications. As a UAV would only cover a fraction of the region of interest, the entire region needs to be covered by several UAVs where each UAV accomplishes its own tasks. For the covering of the target region, a working method consisting of three levels has been developed. The initial step employs the Voronoi partition technique to create a number of convex sub-polygonal areas inside the target area. In the second level, each sub-polygonal area is partitioned to provide a near-optimal collection of waypoints. At the third and final level, we find a path that visits each of the waypoints without colliding with anything and is as short as feasible. Collision due to both static and dynamic obstacles is also considered for avoidance. The first and second-level partitioning processes are carried out offline, whereas path planning is handled in real-time. Traditional methods like Particle Swarm Optimisation (PSO), Genetic Algorithm (GA), and Cuckoo Optimisation Algorithm (COA) are used to evaluate the proposed work. The evaluated result of the proposed work is compared with the existing work on area coverage in terms of the percentage of inside and outside area coverage. Also, the performance of the proposed dynamic path planning method is compared with TSP and the Improved Follow the Gap Method (FGM-I). The outcome demonstrates that the proposed work is far more effective than the existing result and is suitable for the application in issue, which involves area coverage with the presence of obstacles.