UAV Path Planning with Parallel Genetic Algorithms on CUDA Architecture

Abstract

In recent years, Unmanned Aerial Vehicles (UAVs) have been emerged as an attractive technology for different types of military and civil applications, which have gained importance in academic researches. In these emerging research areas, UAV autonomy gets a great part of the study, and mainly it refers the ability for automatic take-off, landing and path planning of UAVs. In this paper, we focused of the path planning of UAVs for controlling a number of waypoints in the mission area. If the area is large and the number of points that must be checked is greater, then it is not possible to check all possible solutions, therefore, we have to use some efficient algorithms, like genetic algorithms (GAs), to calculate an acceptable path. However, if the number of waypoints exceeds a certain number, then we have to use some additional accelerating mechanisms to speed up the calculation time. Typically two techniques are used for speeding up: parallelization and distribution of calculation. In this paper genetic algorithm is parallelized on CUDA architecture by using Graphical Processing Units (GPUs). Experimental results showed that this approach produces efficient solutions in a short time.