This paper describes a scalable method for parallelizing sampling-based motion planning algorithms. It subdivides configuration space (C-space) into (possibly overlapping) regions and independently, in parallel, uses standard (sequential) sampling-based planners to construct roadmaps in each region. Next, in parallel, regional roadmaps in adjacent regions are connected to form a global roadmap. By subdividing the space and restricting the locality of connection attempts, we reduce the work and inter-processor communication associated with nearest neighbor calculation, a critical bottleneck for scalability in existing parallel motion planning methods. We show that our method is general enough to handle a variety of planning schemes, including the widely used Probabilistic Roadmap (PRM) and Rapidly-exploring Random Trees (RRT) algorithms. We compare our approach to two other existing parallel algorithms and demonstrate that our approach achieves better and more scalable performance. Our approach achieves almost linear scalability on a 2400 core LINUX cluster and on a 153,216 core Cray XE6 petascale machine.

In this work, we describe an approach for modeling and simulating group behaviors for pursuit-evasion that uses a graph-based representation of the environment and integrates multi-agent simulation with roadmap-based path planning. Our approach can be applied to more realistic scenarios than are typically studied in most previous work, including agents moving in 3D environments such as terrains, multi-story buildings, and dynamic environments. We also support more realistic three-dimensional visibility computations that allow evading agents to hide in crowds or behind hills. We demonstrate the utility of this approach on mobile robots and in simulation for a variety of scenarios including pursuit-evasion and tag on terrains, in multi-level buildings, and in crowds.