Global Path Planning for Single and Multi-Robot Systems

Global path planning is considered as a fundamental problem for mobile robots. This master project is split into two parts. In the first part we investigate the capabilities of Genetic Algorithms (GA) for solving the global path planning problem in large-scale grid maps. First, we propose a GA approach for efficiently finding an (or near) optimal path in the grid map. We carefully designed GA operators to optimize the search process. We also conduct a comprehensive statistical evaluation of the proposed GA approach in terms of solution quality, and we compare it against the well-known A* algorithm as a reference. Extensive simulation results show that GA is able to find the optimal paths in large environments equally to A* in almost all the simulated cases. Second part of this project deal with the Multi-Robot Path Planning (MRPP) problem. We present two different approaches for collision avoidance in multi-robot systems. First approach is based on prioritized planning, and the second is a new approach that we named it local repair. The preliminary results of the simulation study for the two approaches show that the prioritized planning clearly outperform the local repair approach. In our future work, we will investigate the reasons of that, and will explore different ways to improve it.