Soft subdivision motion planning for complex planar robots

The design and implementation of theoretically-sound robot motion planning algorithms is challenging. Within the framework resolution-exact algorithms, it possible to exploit soft predicates for collision detection. a balancing act between easily implementable their accuracy/effectivity. In this paper, we focus on class planar polygonal rigid robots with arbitrarily complex geometry. We remarkable decomposability property collision-detection such robots. introduce general technique produce decomposition. If an m-gon, complexity approach scales linearly in m. This contrasts O(m^3) known exact planners. It follows that can now routinely any robot. results planners within Soft Subdivision Search (SSS) framework. significant advancement theory sound complete implemented decomposed our open-source Core Library. experiments show are effective, perform real time non-trivial environments, outperform many sampling-based methods.