Abstract: Full coverage path planning is one of the important contents of unmanned system path planning. With the continuous development of Unmanned Aerial Vehicle(UAV) technology, UAV full coverage path planning has been importantly used in many fields. However, in this process, no-fly zones and obstacles often appear, and path planning needs to be carried out to ensure flight safety and flight efficiency. Based on convex partition optimization, this paper proposes a full coverage path planning method for complex plots with complex obstacles, which reduces the length of the coverage path and the total time of the coverage path. Complex plots often contain smooth curves or rugged inner-concave boundary contours. This article first uses an improved Douglas-Peucker algorithm to compress complex plot boundaries into complex polygonal boundaries, and then uses convex-concave points to test the vertices. Then, the shortest main line direction is identified by rotating the main line, and then Probabilistic Roadmaps(PRM) is used to find the shortest auxiliary line. Four convex division strategies are employed to optimize the convex division of the complex land, so that the UAV can be in the full coverage process, with shorter middle path and higher work efficiency. Finally, computer simulations on the test plots are performed to achieve the overall path ratio of 67.6% and 54.9%. The proposed method is compared with other convex partition optimization algorithms for the same map, showing that our method outperforms other methods in terms of path length and planning time.

Key words: complete coverage path planning, multi-UAVs, complex plot, complex obstacle, convex partition optimization, improved Douglas-Peucker algorithm, probabilistic roadmaps method