Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach

Luca Furieri, Clara Lucía Galimberti, Muhammad Zakwan, Giancarlo Ferrari-Trecate
Proceedings of The 4th Annual Learning for Dynamics and Control Conference, PMLR 168:571-583, 2022.

Abstract

Large-scale cyber-physical systems require that control policies are distributed, that is, that they only rely on local real-time measurements and communication with neighboring agents. Optimal Distributed Control (ODC) problems are, however, highly intractable even in seemingly simple cases. Recent work has thus proposed training Neural Network (NN) distributed controllers. A main challenge of NN controllers is that they are not dependable during and after training, that is, the closed-loop system may be unstable, and the training may fail due to vanishing gradients. In this paper, we address these issues for networks of nonlinear port-Hamiltonian (pH) systems, whose modeling power ranges from energy systems to non-holonomic vehicles and chemical reactions. Specifically, we embrace the compositional properties of pH systems to characterize deep Hamiltonian control policies with built-in closed-loop stability guarantees – irrespective of the interconnection topology and the chosen NN parameters. Furthermore, our setup enables leveraging recent results on well-behaved neural ODEs to prevent the phenomenon of vanishing gradients by design. Numerical experiments corroborate the dependability of the proposed architecture, while matching the performance of general neural network policies.

Cite this Paper

BibTeX
@InProceedings{pmlr-v168-furieri22a, title = {Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach}, author = {Furieri, Luca and Galimberti, Clara Luc\'ia and Zakwan, Muhammad and Ferrari-Trecate, Giancarlo}, booktitle = {Proceedings of The 4th Annual Learning for Dynamics and Control Conference}, pages = {571--583}, year = {2022}, editor = {Firoozi, Roya and Mehr, Negar and Yel, Esen and Antonova, Rika and Bohg, Jeannette and Schwager, Mac and Kochenderfer, Mykel}, volume = {168}, series = {Proceedings of Machine Learning Research}, month = {23--24 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v168/furieri22a/furieri22a.pdf}, url = {https://proceedings.mlr.press/v168/furieri22a.html}, abstract = {Large-scale cyber-physical systems require that control policies are distributed, that is, that they only rely on local real-time measurements and communication with neighboring agents. Optimal Distributed Control (ODC) problems are, however, highly intractable even in seemingly simple cases. Recent work has thus proposed training Neural Network (NN) distributed controllers. A main challenge of NN controllers is that they are not dependable during and after training, that is, the closed-loop system may be unstable, and the training may fail due to vanishing gradients. In this paper, we address these issues for networks of nonlinear port-Hamiltonian (pH) systems, whose modeling power ranges from energy systems to non-holonomic vehicles and chemical reactions. Specifically, we embrace the compositional properties of pH systems to characterize deep Hamiltonian control policies with built-in closed-loop stability guarantees – irrespective of the interconnection topology and the chosen NN parameters. Furthermore, our setup enables leveraging recent results on well-behaved neural ODEs to prevent the phenomenon of vanishing gradients by design. Numerical experiments corroborate the dependability of the proposed architecture, while matching the performance of general neural network policies.} }
Endnote
%0 Conference Paper %T Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach %A Luca Furieri %A Clara Lucía Galimberti %A Muhammad Zakwan %A Giancarlo Ferrari-Trecate %B Proceedings of The 4th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2022 %E Roya Firoozi %E Negar Mehr %E Esen Yel %E Rika Antonova %E Jeannette Bohg %E Mac Schwager %E Mykel Kochenderfer %F pmlr-v168-furieri22a %I PMLR %P 571--583 %U https://proceedings.mlr.press/v168/furieri22a.html %V 168 %X Large-scale cyber-physical systems require that control policies are distributed, that is, that they only rely on local real-time measurements and communication with neighboring agents. Optimal Distributed Control (ODC) problems are, however, highly intractable even in seemingly simple cases. Recent work has thus proposed training Neural Network (NN) distributed controllers. A main challenge of NN controllers is that they are not dependable during and after training, that is, the closed-loop system may be unstable, and the training may fail due to vanishing gradients. In this paper, we address these issues for networks of nonlinear port-Hamiltonian (pH) systems, whose modeling power ranges from energy systems to non-holonomic vehicles and chemical reactions. Specifically, we embrace the compositional properties of pH systems to characterize deep Hamiltonian control policies with built-in closed-loop stability guarantees – irrespective of the interconnection topology and the chosen NN parameters. Furthermore, our setup enables leveraging recent results on well-behaved neural ODEs to prevent the phenomenon of vanishing gradients by design. Numerical experiments corroborate the dependability of the proposed architecture, while matching the performance of general neural network policies.
APA
Furieri, L., Galimberti, C.L., Zakwan, M. & Ferrari-Trecate, G.. (2022). Distributed Neural Network Control with Dependability Guarantees: a Compositional Port-Hamiltonian Approach. Proceedings of The 4th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 168:571-583 Available from https://proceedings.mlr.press/v168/furieri22a.html.

Related Material