American Association for the Advancement of Science, Science Robotics, 26(4), 2019
DOI: 10.1126/scirobotics.aau5872
Full text: Unavailable
A method for learning agile control policies uses simulated data to enable precise, efficient movements in a complex physical robot.