GPU-Accelerated Flight Route Planning for Multi-UAV Systems Using Simulated Annealing

摘要

In recent years, Unmanned Aerial Vehicles (UAVs) have been preferred in different application domains such as border surveillance, firefighting, photography, etc. With the decreasing cost of UAVs, to accomplish the mission quickly, these applications facilitates the usage of multiple UAVs instead of using a single large UAV. This makes the trajectory planning problem of UAVs more complicated. Most of the users get help from the evolutionary algorithms. However, increased complexity of the problem necessitates additional mechanism, such as parallel programming, to speed up the calculation process. Therefore, in this paper, it is aimed to solve the path planning problem of multiple UAVs with parallel simulated annealing algorithms which is executed on parallel computing platform: CUDA. The efficiency and the effectiveness of the proposed parallel SA approach are demonstrated through simulations under different scenarios.