# Interactive Dynamic Walking: Learning Gait Switching Policies with Generalization Guarantees

arXiv: https://arxiv.org/abs/2109.13417 · HTML: https://mlsanity.com/paper/2109.13417/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 32 references that are in the map.
- First posted on arXiv 28 Sep 2021.

## Builds on (most cited first, 10 of 10)

- [Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data](https://mlsanity.com/paper/1703.11008/) (2017)
- [PAC-Bayes with Backprop](https://mlsanity.com/paper/1908.07380/) (2019)
- [PAC-Bayes Control: Learning Policies that Provably Generalize to Novel Environments](https://mlsanity.com/paper/1806.04225/) (2018)
- [Robust Feedback Motion Policy Design Using Reinforcement Learning on a 3D Digit Bipedal Robot](https://mlsanity.com/paper/2103.15309/) (2021)
- [Probably Approximately Correct Vision-Based Planning using Motion Primitives](https://mlsanity.com/paper/2002.12852/) (2020)
- [Toward Safety-Aware Informative Motion Planning for Legged Robots](https://mlsanity.com/paper/2103.14252/) (2021)
- [Global Position Control on Underactuated Bipedal Robots: Step-to-step Dynamics Approximation for Step Planning](https://mlsanity.com/paper/2011.06050/) (2020)
- [Safe Adaptive Switching among Dynamical Movement Primitives: Application to 3D Limit-Cycle Walkers](https://mlsanity.com/paper/1810.00527/) (2018)
- [An Inverse Dynamics Approach to Control Lyapunov Functions](https://mlsanity.com/paper/1910.10824/) (2019)
- [Input-to-State Stability of Periodic Orbits of Systems with Impulse Effects via Poincar'e Analysis](https://mlsanity.com/paper/1712.03291/) (2017)

## Abstract (arXiv)

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.