Predicted Composite Signed-Distance Fields for Real-Time Motion Planning in Dynamic Environments

We present a novel framework for motion planning in dynamic environments that\naccounts for the predicted trajectories of moving objects in the scene. We\nexplore the use of composite signed-distance fields in motion planning and\ndetail how they can be used to generate signed-distance fields (SDFs) in\nreal-time to incorporate predicted obstacle motions. We benchmark our approach\nof using composite SDFs against performing exact SDF calculations on the\nworkspace occupancy grid. Our proposed technique generates predictions\nsubstantially faster and typically exhibits an 81--97% reduction in time for\nsubsequent predictions. We integrate our framework with GPMP2 to demonstrate a\nfull implementation of our approach in real-time, enabling a 7-DoF Panda arm to\nsmoothly avoid a moving robot.\n