Path Planning of UAV Navigation Mark Inspection Using a K-means Clustering ACA

Abstract

With increasing speeds and application of artificial intelligence in the shipping industry, unmanned aerial vehicle (UAV) technology has been applied to navigation mark inspections to improve the inspection efficiency and safety. Aimed at the UAV path planning problem of navigation mark inspection, this paper proposes an improved K-means clustering ant colony algorithm (KCACA) to design the shortest route for UAV navigation mark inspections. First, the K-means algorithm and the UAV maximum flight distance were used to cluster the navigation marks, which were then split into several secondary clusters. Each cluster was regarded as an independent traveling salesman problem to be evaluated using the ant colony algorithm (ACA). Second, after optimizing the ACA pheromone update formula, the attenuation factor was gradually reduced according to the number of iterations. Experiments showed that the improved KCACA not only optimized the shortest path but also calculated the optimal path in a short time, improving the operation efficiency.