Path planning is a subject of wide interest in applications for unmanned aerial vehicles. A great variety of solutions have been generated to find paths free of collisions in three-dimensional scenarios, among them, the algorithm based on independent mobile points MBPF which uses the concept of artificial potential fields for the evasion of obstacles, is an algorithm that can be used in both 2D and 3D environments with static or dynamic obstacles and manages to avoid falling into local minimums. MBPF can find trajectories in a variety of complex 3D scenarios. However, it can be improved so that the trajectory reduces its distance, smooth the route and in some cases, converge. To improve the response of the algorithm, 3 improvements are proposed, one based on a modification to the original algorithm and another two based on a post processing of the result. the first improvement directly modifies the initial distribution of the starting points of the algorithm by densifying the points in the areas close to the obstacles, the two remaining improvements are made by post-processing the trajectory obtained for which Visibility Graphs are used and Active Contours. As a result, a considerable reduction in the distance to be traveled is evidenced in the final trajectory and a greater smoothness is perceived in the curves.
展开▼
