When it comes to throwing things, even humans sometimes struggle. It is even more challenging for robots. Researchers have come up with a framework that teaches robot how to throw objects accurately. It comes with “a nominal tracking policy for the end-effector, a high-frequency residual policy to enhance tracking accuracy, and an optimization-based module to improve end-effector acceleration control.” It achieves a landing error of 0.28m at a range of 6 meters.
As the researchers explain, this robot achieved a success rate of 56.8% hitting small targets randomly placed at distances of 5m.
[HT]