PaRRT: Parallel rapidly exploring random tree (RRT) based on MapReduce

The basic Rapidly Exploring Random Trees (RRT) method is recognized as a very effective solution to resolve motion planning problems. The large scale of data provided by the emerging of smart paradigm makes the running of RRT algorithm a challenging task. In fact, research is increasingly oriented towards the parallelization of RRT. In this paper, we propose a new extension of RRT by using MapReduce, the programming model designed to perform parallel distributed computing on very large datasets. The experimental results demonstrate that our PaRRT can scale well and efficiently large data structures.