Improving the performance of RRT path planning of excavators by embedding heuristic rules
Seied Mohammad Langari, Faridaddin Vahdatikhaki, Amin Hammad
Abstract
Improving the safety and productivity of earthwork operations is of paramount importance, especially in congested sites where collisions are more probable. Real-time Location Systems and Automated Machine Guidance and Control are expected to improve both the safety and productivity of earthwork operations by providing excavator operators with a higher level of support regarding the path planning of excavators based on site conditions. However, in spite of the large number of studies related to automated path planning of excavators using well-established algorithms from robotics, such as Rapidly-exploring Random Trees and Probabilistic Roadmaps, these studies do not fully consider the engineering constraints of the equipment and do not result in smooth and optimal paths that can be applied in practice. This paper aims to develop a more practical algorithm for the path planning of excavators by embedding heuristic rules and engineering constraints specific to excavators. The proposed method is implemented and tested in a game engine environment. The efficiency of the proposed method in generating a collision-free path, which is expected to result in improved productivity, is validated both quantitatively and visually. The comparative results with other recent and modified versions of the RRT algorithm show that the proposed algorithm is able to find a more realistic path in a shorter time.