?? Collision detection is a crucial stage for many #robotic applications including motion planning, trajectory optimization, or simulation. ? Our paper introduces a randomized smoothing approach to compute collision detection derivatives for convex shapes—unlocking faster, more accurate, and efficient solutions. ?? Key Takeaways: ? Differentiable collision detection improves #simulation and planning ? Randomized smoothing captures crucial #gradient information ? Implemented in #HPP-FCL and #Pinocchio for real-world applications ?? Read the full paper with #visuals: https://lnkd.in/dTrm4hdc #AGIMUSproject #paper #AI #eu
Convince Project, PILLAR-Robots, IntelliMan Project, CORESENSE, RegoProject, Sestosenso, euROBIN project,Adra - AI-Data-Robotics-Association, AI-on-Demand Platform CAIRNE MANOLO Project LAAS-CNRS, Inria, PAL Robotics PAL France, Q-PLAN INTERNATIONAL, Airbus, THIMM Obaly k.s., Czech Institute of Informatics, Robotics and Cybernetics, KLEEMANN