The Dynamic Rapidly-exploring Generalized Bur Tree (DRGBT) is a sampling-based motion planner that keeps a robotic manipulator on a collision-free path while obstacles move unpredictably. Its key construction is a dynamic expanded bubble of guaranteed-free configuration space, computed from measured obstacle distances, which supplies sufficient conditions for provably safe motion under the arm's kinematic constraints. A timing analysis and scheduling of the planner's routines let a full replanning cycle run in real time on ordinary sequential CPU hardware without GPUs.
The Dynamic Rapidly-exploring Generalized Bur Tree (DRGBT) is a sampling-based motion planner for robotic manipulators operating in dynamic environments, where obstacles move unpredictably and a plan…