PRM Algoritmasinin Kararli Hale Getirilmesi

Abstract

Robots are mechanical devices programmed to perform certain repetitive functions. They can also be programmed to perform many tasks that can be complex or dangerous for humans. In order for robots to be used more effectively and resources more efficiently, they must cross the distance between the starting points and the target points with the least cost. For this, many path planning algorithms have been developed. In this study, the A ? algorithm, which gives optimum results in terms of distance, and the PRM algorithm, which is one of the algorithms that give results in the fastest time, have been examined in detail, and various tests have been made by offering suggestions to improve the PRM algorithm. In the study, the outputs of the A ? algorithm were accepted as optimum and compared with the other algorithm outputs accordingly. Since the PRM algorithm gives different results each time it is run, in order to stabilize this algorithm, the corner points of the obstacles in the environment were also assigned nodes and this method was tested for different environments. The PRM algorithm working with the new method is named PRM-corner. With this new and original method applied, it has been observed that the PRM algorithm both gives stable results like the A ? algorithm and produces fast results from the A ? algorithm. © 2020 IEEE.