HCI Research: AR Foundation Interaction (in progress)
Direct responsibilities:
Interaction design, System prototyping, Empirical evaluation.
Hands-on contributions:
Interaction models and policies, Empirical evidence on adaptive AR, Design principles for safety-critical AR, Open-source AR prototypes.
About the project:
Research at UCL Interaction Centre exploring human-centered interaction policies for adaptive AR and autonomous systems. The work focuses on how interfaces should regulate their presence, modality, and information density based on human attention, context, and system uncertainty in safety-critical environments.
Abstract
Augmented Reality (AR) and Artificial Intelligence (AI) are increasingly embedded in systems that sense, decide, and act in the physical world, including autonomous vehicles, robots, teleoperation, and safety-critical decision tools. In these scenarios, interaction design is no longer only about presenting information, but about regulating attention, trust, and control under different contexts.
Recent HCI research has started to move away from static, visually dominant interfaces, and instead asks how interactive systems can better adapt to human attention, understanding, and changing levels of autonomy. For example, interfaces that track user attention can respond to what users have actually noticed, rather than what the system assumes they have seen. Effective human–AI collaboration relies on clearly defined roles, especially when system behaviour is uncertain. Similarly, work on shared control shows that increasing autonomy changes both cognitive load and user trust, even when task performance remains stable.
The core problem this research addresses is that there is no human-centered framework for how adaptive AR systems should decide when to intervene, how to communicate, and when to remain silent, particularly in systems where humans and automation share control.
By grounding AR interaction design in human-centered metrics and adaptive decision making, this research aims to support safer and more trustworthy integration of AR into autonomous systems. The outcomes are relevant to HCI, AR, human–robot interaction, and safety-critical system design, and are intended to contribute to top-tier venues such as CHI.