Lie Group Forced Variational Integrator Networks for Learning and Control of Robot Systems

Valentin Duruisseaux

Thai Duong

Melvin Leok

Nikolay Atanasov

Department of Mathematics,
University of California, San Diego

Department of Electrical and Computer Engineering,
University of California, San Diego

[Extended Paper]

[Code]

Incorporating prior knowledge of physics laws and structural properties of dynamical systems into the design of deep learning architectures has proven to be a powerful technique for improving their computational efficiency and generalization capacity. Learning accurate models of robot dynamics is critical for safe and stable control. Autonomous mobile robots, including wheeled, aerial, and underwater vehicles, can be modeled as controlled Lagrangian or Hamiltonian rigid-body systems evolving on matrix Lie groups. In this paper, we introduce a new structure-preserving deep learning architecture, the Lie group Forced Variational Integrator Network (LieFVIN), capable of learning controlled Lagrangian or Hamiltonian dynamics on Lie groups, either from position-velocity or position-only data. By design, LieFVINs preserve both the Lie group structure on which the dynamics evolve and the symplectic structure underlying the Hamiltonian or Lagrangian systems of interest. The proposed architecture learns surrogate discrete-time flow maps allowing accurate and fast prediction without numerical-integrator, neural-ODE, or adjoint techniques, which are needed for vector fields. Furthermore, the learnt discrete-time dynamics can be utilized with computationally scalable discrete-time (optimal) control strategies.

Paper

Thai Duong, Nikolay Atanasov

Lie Group Forced Variational Integrator Networks for Learning and Control of Robot Systems

Under review, 2022.

[pdf]

Code

[github]

Citation

If you find our papers/code useful for your research, please cite our work as follows.

1. T. Duong, N. Atanasov. Lie Group Forced Variational Integrator Networks for Learning and Control of Robot Systems. arxiv, 2022.

@article{duruisseaux2022lie,
                                  

 title={Lie Group Forced Variational Integrator Networks for Learning and Control of Robot Systems},
                                  

 author={Duruisseaux, Valentin and Duong, Thai and Leok, Melvin and Atanasov, Nikolay},
                                  

 journal={arXiv preprint arXiv:2211.16006},
                                  

 year={2022}
}

Acknowledgements

We gratefully acknowledge support from NSF CCF-2112665, NSF DMS-1345013, NSF DMS-1813635, and AFOSR FA9550-18-1-0288.
This webpage template was borrowed from https://akanazawa.github.io/cmr/.