Beschreibung
This paper presents a comparison of multiple global path planning algorithms regarding their usage in multi-robot systems. Most performance-based comparisons only focus on computation time and path length, which are not the most important indicators when transporting an object using a multi-robot formation. Therefore, we decided to add the distance to obstacles and curvature of the path as performance metrics to support the selection of a fitting global path planner for a multi-robot application. In the comparison, we included the PRM, RRT*, Relaxed A* and Voronoi path planning algorithms. Each algorithm performs a specified number of planning tasks and is evaluated by the planning success, path length, obstacle distance and curvature, which are visualized below.
