GPU Parallelization of Policy Iteration RRT#

Sampling-based planning has become a de facto standard for complex robots given its superior ability to rapidly explore high-dimensional configuration spaces. Most existing optimal sampling-based planning algorithms are sequential in nature and cannot take advantage of wide parallelism available on modern computer hardware. Further, tight synchronization of exploration and exploitation phases in these algorithms limits sample throughput and planner performance. Policy Iteration RRT# (PI-RRT#) exposes fine-grained parallelism during the exploitation phase, but this parallelism has not yet been evaluated using a concrete implementation. We first present a novel GPU implementation of PI-RRT#'s exploitation phase and discuss data structure considerations to maximize parallel performance. Our implementation achieves 3-4x speedup over a serial PI-RRT# implementation for a 77.9% decrease in overall planning time on average. As a second contribution, we introduce the Batched-Extension RRT# algorithm, which loosens the synchronization present in PI-RRT# to realize independent 12.97x and 12.54x speedups under serial and parallel exploitation, respectively.

[1]  Oliver Brock,et al.  Balancing Exploration and Exploitation in Sampling-Based Motion Planning , 2014, IEEE Transactions on Robotics.

[2]  Marco Pavone,et al.  Fast marching tree: A fast marching sampling-based method for optimal motion planning in many dimensions , 2013, ISRR.

[3]  Panagiotis Tsiotras,et al.  Incremental sampling-based motion planners using policy iteration methods , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[4]  Leslie Pack Kaelbling,et al.  Hierarchical task and motion planning in the now , 2011, 2011 IEEE International Conference on Robotics and Automation.

[5]  B. Faverjon,et al.  Probabilistic Roadmaps for Path Planning in High-Dimensional Con(cid:12)guration Spaces , 1996 .

[6]  Emilio Frazzoli,et al.  Sampling-based algorithms for optimal motion planning , 2011, Int. J. Robotics Res..

[7]  Dinesh Manocha,et al.  GPU-based parallel collision detection for fast motion planning , 2012, Int. J. Robotics Res..

[8]  Siddhartha S. Srinivasa,et al.  Informed RRT*: Optimal sampling-based path planning focused via direct sampling of an admissible ellipsoidal heuristic , 2014, 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[9]  Siddhartha S. Srinivasa,et al.  Batch Informed Trees (BIT*): Sampling-based optimal planning via the heuristically guided search of implicit random geometric graphs , 2014, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[10]  Marco Pavone,et al.  Group Marching Tree: Sampling-Based Approximately Optimal Motion Planning on GPUs , 2017, 2017 First IEEE International Conference on Robotic Computing (IRC).

[11]  Steven M. LaValle,et al.  Planning algorithms , 2006 .

[12]  Nathan Bell,et al.  Thrust: A Productivity-Oriented Library for CUDA , 2012 .

[13]  Emilio Frazzoli,et al.  Massively parallelizing the RRT and the RRT , 2011, 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[14]  John D. Owens,et al.  Gunrock , 2017, ACM Trans. Parallel Comput..

[15]  Milan Simic,et al.  Sampling-Based Robot Motion Planning: A Review , 2014, IEEE Access.

[16]  S. LaValle Rapidly-exploring random trees : a new tool for path planning , 1998 .

[17]  Lydia E. Kavraki,et al.  Probabilistic roadmaps for path planning in high-dimensional configuration spaces , 1996, IEEE Trans. Robotics Autom..

[18]  Nikolaus Correll,et al.  C-FOREST: Parallel Shortest Path Planning With Superlinear Speedup , 2013, IEEE Transactions on Robotics.

[19]  Ron Alterovitz,et al.  Parallel sampling-based motion planning with superlinear speedup , 2012, 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[20]  Panagiotis Tsiotras,et al.  Use of relaxation methods in sampling-based algorithms for optimal motion planning , 2013, 2013 IEEE International Conference on Robotics and Automation.