Supervisory control via AR for teleoperation under communication delays: an in-space assembly use case

Stavropoulos, Ioannis Marios ORCID: https://orcid.org/0009-0000-7161-7479, Kucukyilmaz, Ayse, Martinez, Carol, Finnegan, Daniel J. ORCID: https://orcid.org/0000-0003-1169-2842 and Hernandez, Juan David ORCID: https://orcid.org/0000-0002-9593-6789 2026. Supervisory control via AR for teleoperation under communication delays: an in-space assembly use case. Presented at: 18th International Conference on Social Robotics (ICSR+ART 2026), London, UK, 1-4 July 2026. Proceedings of the 18th International Conference on Social Robotics (ICSR+ART 2026). Lecture Notes in Artificial Intelligence Cham, Switzerland: Springer, Item availability restricted.

	PDF - Accepted Post-Print Version Restricted to Repository staff only Download (5MB)
	PDF (Provisional file) - Accepted Post-Print Version Download (17kB)

Abstract

Rigorous safety standards in space robotics require human-in-the-loop operation, thus limiting the level of robotic autonomy that can be deployed in practice. While shared control approaches are effective in low-latency settings, high communication delays force operators into inefficient \textit{move-and-wait} strategies. Supervisory control has therefore emerged as a promising alternative for distant space missions like station keeping duties and assembly. However, existing solutions often rely on non-scalable task representations, such as Finite State Machines (FSMs) or provide immersive interfaces primarily designed for low-level robot control. As a result, they offer limited support for specifying complex tasks and managing execution failures in a structured and recoverable way. In this work, we present an AR-supported supervisory control system for in-space assembly tasks that allows operators to specify assembly goals directly within a 3D AR environment and executes them using \acp{BT}. BTs enable failure detection and recovery, while the AR interface provides visual feedback when failures occur. We evaluate the proposed approach in a user study comparing supervisory control with a shared control baseline under two communication latency conditions ($0s$ and $2s$ delay). Our results indicate that supervisory control significantly reduces task completion time, operator effort and workload, while improving robotic motion efficiency and usability. Furthermore, supervisory control remains robust under communication delay, whereas shared control performance degrades.

Item Type:	Conference or Workshop Item - published (Paper)
Status:	In Press
Schools:	Schools > Computer Science & Informatics
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Uncontrolled Keywords:	Supervisory control,Shared Control,Space Robotics,Augmented Reality (AR),Behavior Trees (BTs),Failure Recovery
Publisher:	Springer
Date of First Compliant Deposit:	17 May 2026
Date of Acceptance:	April 2026
Last Modified:	02 Aug 2026 07:36
URI:	https://orca.cardiff.ac.uk/id/eprint/187027

Actions (repository staff only)

Edit Item