Adaptive Mimic: Deep Reinforcement Learning of Parameterized Bipedal Walking from Infeasible References
This work addresses efficient robot locomotion learning for robotics by reducing reliance on laboriously tuned references, though it is incremental in improving imitation learning methods.
The paper tackles the problem of learning parameterized bipedal walking in deep reinforcement learning by proposing an adaptive reward function for imitation learning from references, enabling the agent to mimic low-quality or infeasible references to expedite learning and achieve high performance beyond the references.
Not until recently, robust robot locomotion has been achieved by deep reinforcement learning (DRL). However, for efficient learning of parametrized bipedal walking, developed references are usually required, limiting the performance to that of the references. In this paper, we propose to design an adaptive reward function for imitation learning from the references. The agent is encouraged to mimic the references when its performance is low, while to pursue high performance when it reaches the limit of references. We further demonstrate that developed references can be replaced by low-quality references that are generated without laborious tuning and infeasible to deploy by themselves, as long as they can provide a priori knowledge to expedite the learning process.