Usefulness of using Nvidia IsaacSim and IsaacGym for AI robot manipulation training

摘要

AI technology is increasingly being used to learn robotic behavior. Various types of learning techniques are being applied to learn optimal robotics control or autonomous driving of driverless cars, and attempts are being made to use reinforcement learning techniques to learn robot behavior. When building data for robot behavior learning is not practical in terms of cost and time, imitation learning and reinforcement learning become realistic options for behavior learning using simulators. In this paper, we introduce a robot learning method using Nvidia’s IsaacSim and IsaacGym, and examine the performance result required for learning to test its feasibility.