Neural Collision Clearance Estimator for Batched Motion Planning скачать в хорошем качестве

Neural Collision Clearance Estimator for Batched Motion Planning

Neural Collision Clearance Estimator for Batched Motion Planning By J. Chase Kew, Brian Ichter, Maryam Bandari, Tsang-Wei Edward Lee, Aleksandra Faust Abstract: We present a neural network collision checking heuristic, ClearanceNet, and a planning algorithm, CN-RRT. ClearanceNet predicts separation distance with respect to a workspace. CN-RRT efficiently computes a motion plan by leveraging three key features of ClearanceNet. CN-RRT explores the space by expanding multiple nodes at the same time, adaptively relaxes its clearance requirements for more difficult problems, shifts its nodes in the direction of ClearanceNet’s gradient and repairs any residual errors with a traditional RRT, maintaining theoretical probabilistic completeness guarantees. In configuration spaces with up to 30 degrees of freedom, ClearanceNet achieves 845x speedup over traditional collision detection methods, while CN-RRT accelerates motion planning by up to 42% and finds paths up to 36% more efficient. Keywords: Motion planning, machine learning, robotics