Multi-UAV Path Planning with Parallel Genetic Algorithms on CUDA Architecture

Abstract

In recent years, the use of Unmanned Aerial Vehicles (UAVs) has grown quickly due to its low cost and easily programming for autonomous path following for accomplishing different types of missions. Due to the numerous advantages of multi-UAVs, when comparing with a single powerful one, to perform reconnaissance, monitoring, detection and surveying missions the use of multi-UAVs is generally preferred. While the number of control points and the number of UAVs are increased, the complexity of the problem also increases. This paper presents a solution to the problem of minimum time coverage of ground areas using a number of UAVs. The solution is divided into two parts: Firstly the area is partitioned with K-means clustering and then the problem is solved in each cluster with parallel genetic algorithm approach on CUDA architecture. To illustrate the methodology, the paper presents the experimental results obtained with a multi-UAV system, which has a different number of control points. The results showed the proposed approach produces efficient solutions for these type NP-Hard problems of homeland security applications like wide-area surveillance and site security by using multiple UAVs.