Settings where multiple AR users are co-located present unique challenges and opportunities: supporting rich AR experiences can overwhelm both the wireless network and the edge-computing infrastructure, yet co-location can also enhance the robustness of environmental-awareness algorithms and enable truly engaging collaborative experiences. With support from the National Science Foundation, Cisco, and NASA, the Duke I3T Lab has been developing techniques that reduce the resource consumption of co-located AR devices, exploit data from earlier sessions to accelerate later ones, and intelligently fuse the streams collected by multiple headsets to improve spatial and semantic awareness.
A specific type of multi-agent scenario that involves AR occurs when the co-located entities are a user wearing an AR headset and a robot working alongside that user. In this scenario, AR-based situated analytics can provide real-time information about the state of the robot and facilitate robot control, improving the user’s task performance and trust in their robotic collaborator. We are developing a number of solutions for AR-mediated human-robot collaborations in our ongoing work. In May 2025, we presented two demonstrations of AR-mediated human-robot collaboration at the 2025 Athena AI Institute Showcase.
Recent and selected publications
[TOSN22] G. Lan, Z. Liu, Y. Zhang, T. Scargill, J. Stojkovic, C. Joe-Wong, and M. Gorlatova, Edge-assisted Collaborative Image Recognition for Mobile Augmented Reality, ACM Transactions on Sensor Networks, Vol. 18, No 1, Feb. 2022.
[IoTJ22b] Y. Han, Y. Chen, R. Wang, J. Wu, and M. Gorlatova, Intelli-AR Preloading: A Learning Approach to Proactive Hologram Transmissions in Mobile AR, IEEE Internet of Things Journal, Vol. 9, No. 18, Sept. 2022.
[CoNext20] X. Ran, C. Slocum, Y.-Z. Tsai, K. Apicharttrisorn, M. Gorlatova, and J. Chen, Multi-User Augmented Reality with Communication Efficient and Spatially Consistent Virtual Objects, in Proc. ACM CoNEXT, Dec. 2020.
Great thanks to our sponsors!
This research is supported by the awards from Cisco, NASA, and the National Science Foundation.