Development of an Adaptive RRT Algorithm for Efficient Path Planning of High DOF Dual-arm Robots

This paper propose an adaptive Rapidly-exploring Random Tree (adaptive RRT) for highdimensional path planning. Before the planner extends the RRT, it selects the bodies of robot depending on complexity of extend operation. For instance, if an obstacle is close to the current node on the RRT, then it selects more bodies to increase the dimension of search space. Otherwise, the planner tries to extend the RRT within the lowest dimensional configuration space. We implement our algorithm for a path planning of 19 DOF robot, AMIRO. The results from our experiments show that the adaptive RRT-based path planner is faster than the basic RRT-based path planner