A recursive approach to roadmap-based path planning

Many techniques have been developed to plan paths for mobile robots. The need to provide reasonable paths for mobile robots is critical in the development of useful robots. Many researchers have dismissed roadmap approaches to path planning because of computational inefficiencies and scalability challenges; it is assumed that crude paths result if the problem must be solved in reasonable time. This paper proposes a new approach to path planning that uses recursion to overcome the limitations of other roadmap methods. The JARB algorithm is highly optimized for speed in sparse environments while also providing smooth paths in densely populated environments. By joining recursive branching with the ability to ignore irrelevant obstacles, high quality paths can be produced with reduced computation. We describe the algorithm in detail and present experimental results.