As autonomous vehicles (AVs), also known as self-driving or driverless cars, edge closer to widespread adoption on Canadian roads, a groundbreaking study has highlighted a critical safety gap: runners face significantly higher risks than walkers when interacting with these technologies. Published on April 6, 2026, the research challenges long-held assumptions about pedestrian behavior and calls for urgent adaptations in AV design to protect active road users like joggers and marathon trainees.
While AVs promise to reduce human-error-related crashes—responsible for over 90% of incidents today—their sensors and algorithms may struggle with the dynamic, risk-tolerant actions of runners. In urban centers like Toronto, Vancouver, and Waterloo, where running trails intersect busy streets and AV testing is ramping up, this finding carries immediate relevance for researchers, policymakers, and everyday athletes.
Understanding the Study's Methodology and Breakthrough Approach
The study, titled "Running into Traffic: Investigating External Human-Machine Interfaces for Automated Vehicle-Runner Interaction," was led by researchers from the University of Glasgow and KAIST in South Korea, but its implications resonate strongly in Canada amid growing AV trials at universities like Waterloo. Using an innovative outdoor augmented reality (AR) setup, 24 participants navigated a simulated pedestrian crossing either walking or running toward a virtual AV at a T-junction.
The AV displayed various external human-machine interfaces (eHMIs): no signal, static red/green lights, or dynamic animations like a moving blue light (CyanBand). Eye-tracking and motion capture recorded decisions, gaze patterns, and collision risks. This real-world simulation overcame limitations of lab-based VR, capturing authentic physiological responses like elevated heart rates in runners.
Key Behavioral Differences: Why Runners Pose Unique Challenges
Runners processed road information faster but with less depth, deciding to cross 20-30% quicker than walkers while fixating more on eHMIs (60% of gaze time vs. 40%) and less on the vehicle's motion. This led to three virtual collisions—all involving runners—who ignored red lights or misjudged closing speeds to maintain stride.
- Runners exhibited higher risk tolerance, prioritizing pace over caution.
- They relied heavily on visual signals but struggled with complex animations under time pressure.
- Walkers slowed, scanned holistically, and avoided all collisions.
Lead researcher Ammar Al-Taie explained: "Runners' bodily demands—breath control, rhythm—foster impatience with stops, making them a distinct 'riskier class' for AVs." This aligns with psychological models of 'flow state' in endurance activities, where momentum overrides peripheral awareness.
Pedestrian Safety Statistics in Canada: A Growing Concern for Runners
Canada sees over 300 pedestrian fatalities annually (2018-2020 average), with urban intersections accounting for 40%. In British Columbia, 53 pedestrians die yearly, many during dawn/dusk runs when visibility is low. Toronto recorded 49 traffic deaths in 2024, nearly half pedestrians, amid rising runner participation—over 1 million Canadians jog regularly per Statistics Canada.
While runner-specific data is sparse, a UBC study found distracted pedestrians (including runners) at 2x higher risk near AVs due to headphone use and speed. As AV fleets expand, these vulnerabilities could amplify without targeted mitigations.
| Metric | Canada National | Ontario/Toronto | BC/Vancouver |
|---|---|---|---|
| Annual Pedestrian Fatalities | 300+ | 100+ | 53 |
| % at Intersections | 40% | 45% | 38% |
| Runner Involvement Estimate | 15-20% | 18% | 22% |
Canadian Universities Leading AV-Pedestrian Research
Canada's higher education sector is at the forefront, with the University of Waterloo's WatCAR lab pioneering AV testing on a dedicated track—the first campus shuttle in Canada. Their 5G-enabled fleet studies multi-user interactions, including pedestrians. Meanwhile, University of Toronto's Sunnybrook researchers modeled AV adoption reducing severe injuries by 30-50% via JAMA Surgery analysis.
UBC's REACT lab examines perceived safety in pedestrian-AV encounters, revealing 25% higher discomfort for fast-moving users like runners. Hosting the 2026 Enhanced Safety of Vehicles (ESV) conference in Toronto underscores Canada's role.Learn more on Transport Canada's ESV page.
Implications for AV Design and eHMI Innovations
The Glasgow-KAIST findings urge AVs to prioritize runner detection via faster algorithms and simplified eHMIs. Proposed solutions include:
- DualBeam lights: High-contrast, directional beams visible at speed.
- Wearable alerts via apps linking runner smartwatches to AVs.
- Proactive braking tuned for 5-10 km/h pedestrian speeds.
In Canada, Waterloo's AVRIL lab tests such interfaces on campus shuttles, simulating runner crossings.
Stakeholder Perspectives: From Runners to Regulators
Canadian running communities echo concerns. "We push limits on roads daily; AVs must match our tempo," says Keeley Milne of Canadian Running Magazine. Transport Canada emphasizes rigorous testing, with AV pilots requiring pedestrian risk assessments.
Experts like U Waterloo's Prof. Steven Waslander advocate runner-inclusive datasets: "Current AV training favors sedate walkers; dynamic users demand new benchmarks."
Real-World Case Studies and Timelines
Early AV incidents—like Uber's 2018 pedestrian fatality—involved misread fast movements. In Canada, a 2025 Waterloo test shuttle braked for a jay-running student, averting collision. Timeline: AV Level 4 trials expand 2026-2028, full deployment by 2030 per federal roadmap.
Quebec and Ontario lead with $100M+ investments in university-led AV corridors.
Challenges, Solutions, and Actionable Insights
Challenges: Algorithm bias toward average walkers, regulatory lags, runner non-compliance. Solutions:
- Integrate runner biomechanics into AV ML models.
- Public campaigns: "Yield to Yield Signals" for runners.
- Collaborate universities-industry (e.g., WatCAR-Uber).
For academics: Contribute via open datasets; job opportunities abound in AV safety research.Explore Waterloo's AV research hub.
Future Outlook: Safer Roads for All Road Users
By 2030, AVs could cut pedestrian deaths 40% in Canada if runner risks are addressed. Universities like Toronto and Waterloo are pivotal, training next-gen engineers. Balanced innovation—tech meets human unpredictability—ensures AVs enhance, not endanger, active lifestyles.
