Publications
Two CHI Honourable Mentions, an IJHCS journal article, ACM VRST, and the earlier work that built the foundation. Full citation metrics on Google Scholar.
28th International Conference on Human-Computer Interaction (HCII 2026) · Montréal, Canada
How visual design choices in shared spaces transform pedestrian experience in pedestrian–AV interaction. Complements SPIU work by studying the surrounding environment, not just the signaling unit.
Forthcoming · July 2026
ACM CHI 2026 · Barcelona, Spain
In-the-wild evaluation (N=21) of a mobile SPIU with integrated cameras and LED displays delivering explicit "WALK"/"STOP" cues. Four-factor design (CarBehavior × Mobility × eHMI × SPIU) shows improvements in understandability, trust, and perceived safety; strongest results with SPIU + vehicle eHMI combined.
ACM CHI 2026 · Barcelona, Spain
Parallel field study (N=20 humans vs. N=20 VLM personas) on street-crossing decisions with autonomous vehicles. Maps where VLM personas can substitute for human participants — and where they cannot.
ACM Symposium on Virtual Reality Software and Technology (VRST 2025) · Montréal, Canada
Parallel VR study (N=40 across Japan and Norway, 32 trials each) across four high-risk scenarios: four-way intersections, mixed traffic with delivery robots, blindspots, and nighttime crossings. SPIU reduces pedestrian decision time 40–80% depending on scenario and cultural context. Open-source Smartpole-VR-AWSIM framework released.
International Journal of Human–Computer Studies, vol. 205, p. 103628, 2025
Group design study (N=25, ages 18–40) compared against three multimodal LLMs (ChatGPT-4, Gemini-Pro, Claude 3.5 Sonnet) as "experts". 90% of participants preferred SPIU-mediated multi-AV interaction over conventional eHMI. LLMs used safety-related keywords 30% more than humans, suggesting a more structured but less nuanced perspective.
International Conference on Intelligent Computing and Emerging Applications (ICEA 2024) · Tokyo, Japan
Expert survey (N=12 automotive UI designers) on eHMI, SPIU, and their combination across five parameters: safety, seamlessness, adaptability, accessibility, and trust. SPIU rated highest on four; combined eHMI+SPIU rated above either alone.
ACM AutomotiveUI 2024 (Adjunct, Work in Progress) · Stanford, USA
Think-tank study with 25 participants brainstorming design solutions for pedestrian–AV interaction emphasising effectiveness, simplicity, visibility, and clarity. Establishes design principles for the integrated SPIU + vehicle eHMI prototypes.
Electronics, vol. 12, no. 20, p. 4207, 2023
Formal concept validation through immersive VR experiments. SPIU combined with vehicle eHMI yields a 21% improvement in pedestrian response time. Significant in both one-way and two-way crossing scenarios (p = 0.001).
IEEE 9th World Forum on Internet of Things (WF-IoT 2023) · Aveiro, Portugal
The foundational SPIU paper. Introduces the idea of an infrastructure-side mediator to reduce pedestrian cognitive load in shared spaces. VR-based proof of concept showing faster and safer decision-making in both single-AV and multi-AV scenarios.