AI “Perceptive Assistance” During Teleoperations of Rovers in Deep Space
Joint PhD research project funded through MSCA COFUND SEED
Joint doctoral research · 2024–2027
Designing perceptive assistance that lets people act with greater awareness when the environment is distant, delayed, and only partially sensed.
Franco-Australian joint doctorateIMT Atlantique × University of AdelaideA CROSSING collaboration linking immersive interaction, human factors, and space-resource research through the MSCA COFUND SEED programme.
Abstract
Future missions to the Moon and Mars will rely on remotely operated rovers for exploration and construction. Extended-Reality interfaces can help operators understand the rover’s environment, but restricted camera views, communication latency, and the absence of natural perceptual or proprioceptive feedback may reduce situational awareness and performance.
This PhD project investigates AI-based “perceptive assistance” for deep-space rover teleoperation. Rather than replacing the operator’s judgement, the research asks how an interface can reveal what is difficult to perceive, make uncertainty legible, and support better-timed action. It combines immersive interaction, human perception, multimodal feedback, state prediction, and human-factors research to design and evaluate assistance adapted to these extreme operating conditions.
Research challenges
- Situational awareness — understand how limited fields of view and non-natural head positions affect the operator’s perception of the remote environment.
- Communication latency — characterize the perceptual and behavioral effects of Earth–Moon delays, including their possible contribution to cybersickness.
- Missing sensory cues — determine which combinations of visual, auditory, haptic, and proprioceptive feedback best compensate for absent or degraded information.
- Adaptive assistance — develop AI agents and state-prediction techniques that select or generate useful assistance according to the task and operating conditions.
Research route
Understand the gap
Model the operator’s perceptual constraints
Controlled Virtual-Reality studies isolate the effects of limited viewpoints, non-natural head positions, delay, and sensory loss on remote action.
Prototype assistance
Connect sensing, prediction, and feedback
Interface concepts combine sensory compensation with AI-based state prediction to help operators interpret what they cannot directly see or feel.
Test in representative conditions
Measure benefits and trade-offs
Analogous scenarios, including underwater teleoperation, create experimentally accessible conditions for assessing performance, situation awareness, and comfort.
Expected impact
The project aims to produce evidence and design principles for intelligible remote-operation interfaces. Beyond space exploration, its results may support teleoperation in other environments affected by degraded communications, limited sensing, or strong physical constraints.
People involved
PhD student
Nazial Kadir
Nazial investigates how immersive interfaces, multimodal feedback, and AI can support operators in deep-space rover teleoperation.
Supervisors
Anna Ma-Wyatt
Brings expertise in active vision, human factors, and human–autonomy teaming, including astronaut–autonomy teaming for space operations.
Research environment
The project draws on the immersive-interaction and eye-tracking facilities of IMT Atlantique and the University of Adelaide. It also benefits from the research environment of the CNRS International Research Laboratory CROSSING and the Andy Thomas Space Resource Centre.
Funding
This research project is funded through SEED, IMT Atlantique’s interdisciplinary, international, and intersectoral doctoral programme co-funded by the European Union under the Horizon Europe Marie Skłodowska-Curie Actions COFUND scheme (grant agreement No. 101126644).