Graph Neural Networks for Multi-Robot
Active Information Acquisition
ICRA 2023 Outstanding paper award in Multi-Robot Systems
Mariliza Tzes *,1, Nikolaos Bousias *,1, Evangelos Chatzipantazis1, George J. Pappas1
1GRASP Lab, University of Pennsylvania
Abstract
This paper addresses the Multi-Robot Active Information Acquisition (AIA) problem, where a team of mobile robots, communicating through an underlying graph, estimates a hidden state expressing a phenomenon of interest. Applica- tions like target tracking, coverage and SLAM can be expressed in this framework. Existing approaches, though, are either not scalable, unable to handle dynamic phenomena or not robust to changes in the communication graph. To counter these shortcomings, we propose an Information-aware Graph Block Network (I-GBNet), an AIA adaptation of Graph Neural Networks, that aggregates information over the graph represen- tation and provides sequential-decision making in a distributed manner. The I-GBNet, trained via imitation learning with a centralized sampling-based expert solver, exhibits permutation equivariance and time invariance, while harnessing the superior scalability, robustness and generalizability to previously unseen environments and robot configurations. Experiments on signifi- cantly larger graphs and dimensionality of the hidden state and more complex environments than those seen in training validate the properties of the proposed architecture and its efficacy the application of localization and tracking of dynamic targets.
Citation:
Tzes, M., Bousias, N., Chatzipantazis, E., & Pappas, G. J. (2023, May). Graph neural networks for multi-robot active information acquisition. In 2023 IEEE International Conference on Robotics and Automation (ICRA) (pp. 3497-3503). IEEE.
Tzes, Mariliza, Nikolaos Bousias, Evangelos Chatzipantazis, and George J. Pappas. "Graph Neural Networks for Multi-Robot Active Information Acquisition." arXiv preprint arXiv:2209.12091 (2022).
Demo Video
3 Drones localize 10 Phone Boxes over Manhattan, New York
More coming soon ...
* With UNITY 3D and Flightmare/Gazebo
Song, Y., Naji, S., Kaufmann, E., Loquercio, A. and Scaramuzza, D., 2021, October. Flightmare: A flexible quadrotor simulator. In Conference on Robot Learning (pp. 1147-1157). PMLR.
Brief Presentation
Find us during ICRA 2023
15:00 - 16:40 | Tue 30 May | PH PODS 31-33 | TuPO2S-14.11
Session: Multi-Robot Systems II (Poster Session)
09:00 - 09:10 | Wed 31 May | Auditorium | WeAT7.1
Session: Award Finalists 3 (Oral)
Questions?
nbousias@seas.upenn.edu : mtzes@seas.upenn.edu
3330 Walnut St, Philadelphia, PA 19104, United States