HYBRID FREE-OBSTACLE PATH PLANNING ALGORITHM USING IMAGE PROCESSING AND GEOMETRIC TECHNIQUES

Abstract

This paper shows a hybrid path planning algorithm based on image processing and geometric techniques, where the main idea is to obtain a free-obstacle path for a mobile robot over a known environment through a camera located in top of it. The algorithm uses image processing operations like dilation, skeletonization, image convolution among others, combined with simple geometric operations like distance between points and middle point calculation. Those operations had the goal to find some amount of navigable points including the starting point to the ending one. The A* algorithm was used to select the shortest combination of navigable points to take the mobile robot through, avoiding all the obstacles in the environment. Some reduction rules were implemented into the proposed algorithm with the purpose of decreasing the number of key points and/or navigable points and thus optimizing the density of the navigable network reducing the total computing time. Finally, the result of testing this approach over a total of 8 different navigation environments are displayed, comparing the number of obstacles and their relationship with the computing time.