CatlNet: Learning Communication and Coordination Policies from CaTL+ Specifications

Wenliang Liu, Kevin Leahy, Zachary Serlin, Calin Belta
Proceedings of The 5th Annual Learning for Dynamics and Control Conference, PMLR 211:705-717, 2023.

Abstract

In this paper, we propose a learning-based framework to simultaneously learn the communication and distributed control policies for a heterogeneous multi-agent system (MAS) under complex mission requirements from Capability Temporal Logic plus (CaTL+) specifications. Both policies are trained, implemented, and deployed using a novel neural network model called CatlNet. Taking advantage of the robustness measure of CaTL+, we train CatlNet centrally to maximize it where network parameters are shared among all agents, allowing CatlNet to scale to large teams easily. CatlNet can then be deployed distributedly. A plan repair algorithm is also introduced to guide CatlNet’s training and improve both training efficiency and the overall performance of CatlNet. The CatlNet approach is tested in simulation and results show that, after training, CatlNet can steer the decentralized MAS system online to satisfy a CaTL+ specification with a high success rate.

Cite this Paper

BibTeX
@InProceedings{pmlr-v211-liu23a, title = {CatlNet: Learning Communication and Coordination Policies from CaTL+ Specifications}, author = {Liu, Wenliang and Leahy, Kevin and Serlin, Zachary and Belta, Calin}, booktitle = {Proceedings of The 5th Annual Learning for Dynamics and Control Conference}, pages = {705--717}, year = {2023}, editor = {Matni, Nikolai and Morari, Manfred and Pappas, George J.}, volume = {211}, series = {Proceedings of Machine Learning Research}, month = {15--16 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v211/liu23a/liu23a.pdf}, url = {https://proceedings.mlr.press/v211/liu23a.html}, abstract = {In this paper, we propose a learning-based framework to simultaneously learn the communication and distributed control policies for a heterogeneous multi-agent system (MAS) under complex mission requirements from Capability Temporal Logic plus (CaTL+) specifications. Both policies are trained, implemented, and deployed using a novel neural network model called CatlNet. Taking advantage of the robustness measure of CaTL+, we train CatlNet centrally to maximize it where network parameters are shared among all agents, allowing CatlNet to scale to large teams easily. CatlNet can then be deployed distributedly. A plan repair algorithm is also introduced to guide CatlNet’s training and improve both training efficiency and the overall performance of CatlNet. The CatlNet approach is tested in simulation and results show that, after training, CatlNet can steer the decentralized MAS system online to satisfy a CaTL+ specification with a high success rate. } }
Endnote
%0 Conference Paper %T CatlNet: Learning Communication and Coordination Policies from CaTL+ Specifications %A Wenliang Liu %A Kevin Leahy %A Zachary Serlin %A Calin Belta %B Proceedings of The 5th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2023 %E Nikolai Matni %E Manfred Morari %E George J. Pappas %F pmlr-v211-liu23a %I PMLR %P 705--717 %U https://proceedings.mlr.press/v211/liu23a.html %V 211 %X In this paper, we propose a learning-based framework to simultaneously learn the communication and distributed control policies for a heterogeneous multi-agent system (MAS) under complex mission requirements from Capability Temporal Logic plus (CaTL+) specifications. Both policies are trained, implemented, and deployed using a novel neural network model called CatlNet. Taking advantage of the robustness measure of CaTL+, we train CatlNet centrally to maximize it where network parameters are shared among all agents, allowing CatlNet to scale to large teams easily. CatlNet can then be deployed distributedly. A plan repair algorithm is also introduced to guide CatlNet’s training and improve both training efficiency and the overall performance of CatlNet. The CatlNet approach is tested in simulation and results show that, after training, CatlNet can steer the decentralized MAS system online to satisfy a CaTL+ specification with a high success rate.
APA
Liu, W., Leahy, K., Serlin, Z. & Belta, C.. (2023). CatlNet: Learning Communication and Coordination Policies from CaTL+ Specifications. Proceedings of The 5th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 211:705-717 Available from https://proceedings.mlr.press/v211/liu23a.html.

Related Material