Adaptive swept volumes generation for human-robot coexistence using Gaussian Processes

摘要: Letting humans and robots share a common space for collaboration is considered consolidated practice. The trajectories followed by the robot must be safe human mate, especially when holds dangerous tools or parts. At same time, productivity preserved, without imposing too restrictive limitations on robot’s movements. This article proposes use of Gaussian Processes to predict motion an operator in robotic cell, with aim controlling speed avoid collisions. An adaptive approach proposed model persistently re-updated. resulting will demonstrated less conservative than previous ones, while at time preserve safety operator. Real experiments have been conducted 7 d.o.f. ABB YuMi robot.