Interactive Dynamic Walking: Learning Gait Switching Policies with Generalization Guarantees
This addresses the challenge of safe and reliable human-robot collaboration in dynamic walking tasks, though it is incremental as it builds on existing DMP and generalization theory methods.
The paper tackles the problem of enabling a bipedal robot to adapt its walking gait to follow a human co-worker based on implicit interaction forces, achieving this by training a neural-network supervisor to switch among Dynamic Movement Primitives with certificates of generalization to novel intentions using PAC-Bayes bounds.
In this paper, we consider the problem of adapting a dynamically walking bipedal robot to follow a leading co-worker while engaging in tasks that require physical interaction. Our approach relies on switching among a family of Dynamic Movement Primitives (DMPs) as governed by a supervisor. We train the supervisor to orchestrate the switching among the DMPs in order to adapt to the leader's intentions, which are only implicitly available in the form of interaction forces. The primary contribution of our approach is its ability to furnish certificates of generalization to novel leader intentions for the trained supervisor. This is achieved by leveraging the Probably Approximately Correct (PAC)-Bayes bounds from generalization theory. We demonstrate the efficacy of our approach by training a neural-network supervisor to adapt the gait of a dynamically walking biped to a leading collaborator whose intended trajectory is not known explicitly.