Robot weightlifting by direct policy search

摘要

This paper describes a method for structuring a robot motor learning task. By designing a suitably parameterized policy, we show that a simple search algorithm, along with biologically motivated con-straints, offers an effective means for motor skill acquisition. The framework makes use of the robot counterparts to several elements found in human motor learning: imitation, equilibrium-point con-trol, motor programs, and synergies. We demon-strate that through learning, coordinated behavior emerges from initial, crude knowledge about a dif-ficult robot weightlifting task. 1