Incremental sampling-based motion planners using policy iteration methods

Recent progress in randomized motion planning has led to the development of a new class of sampling-based algorithms that provide asymptotic optimality guarantees, almost surely, notably the RRT* algorithm among others. The recently proposed RRT# algorithm utilizes dynamic programming ideas (namely, asynchronous value iteration) and implements them incrementally on randomly generated graphs to improve on RRT* and thus obtain high quality solutions with enhanced rates of convergence. In this work we explore a different class of dynamic programming algorithms for solving shortest-path problems on random graphs generated by iterative sampling, which utilize policy iteration instead of value iteration and that are better suited for massive parallelization.