Risk Assessment by a Passenger of an Autonomous Vehicle Among Pedestrians: Relationship Between Subjective and Physiological Measures

Petit, Jeffery; Charron, Camilo; Mars, Franck · 2021 · Crossref

DOI: 10.3389/fnrgo.2021.682119

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates how passengers in autonomous vehicles perceive collision risk when navigating shared spaces with pedestrians, specifically examining the relationship between subjective risk assessments and physiological responses. The research is motivated by the complexity of autonomous navigation in environments lacking road signs, where vehicle trajectories must remain acceptable to both external road users and internal passengers. The authors address the theoretical duality of risk perception—“risk as analysis” (conscious reasoning) and “risk as feelings” (intuitive, affective response)—by testing whether subjective and physiological measures are independent or related systems. The experiment utilized a fixed-base driving simulator with 20 participants who acted as passengers in a fully autonomous vehicle traveling at 30 kph. Participants encountered 32 randomized scenarios involving avoidance maneuvers for pedestrians walking at 4.5 kph. Two key factors were manipulated: the time-to-collision (TTC) at the initiation of the maneuver (2.0, 2.5, 3.0, and 3.5 seconds) and the lateral offset distance between the vehicle and pedestrian (0.5, 1.0, and 1.5 meters). Risk perception was measured using two methods: a real-time subjective assessment via a one-handed analogue potentiometer and electrodermal activity (EDA) recorded via skin conductance sensors on the non-dominant hand. The data were analyzed using hybrid Bayesian networks to determine the conditional dependencies between the manipulated factors and the two risk indicators. The results demonstrated that reducing safety margins—specifically lower TTC values and smaller lateral offsets—increased risk perception according to both subjective and physiological indicators. However, the increase in subjective risk assessment was more pronounced than the corresponding physiological response. Statistical modeling via Bayesian networks revealed that while the two indicators are not redundant, they are also not independent; there is a significant relationship between the subjective and physiological systems of risk perception. The analysis confirmed that both TTC and lateral offset significantly influence the passenger’s perceived risk, validating the hypothesis that vehicle-environment dynamics directly impact passenger comfort and safety perception. The significance of this study lies in its empirical validation of the interdependence between cognitive and affective risk perception systems in autonomous driving contexts. By demonstrating that subjective and physiological measures are linked but distinct, the findings suggest that relying on a single metric may provide an incomplete picture of passenger experience. This has implications for the design of autonomous vehicle algorithms, particularly in shared spaces, where navigation strategies must account for both the logical safety margins and the intuitive comfort zones of passengers to ensure acceptability and trust in automated systems.

Key finding

Reductions in safety margins during pedestrian avoidance maneuvers increase risk perception in both subjective and physiological measures, with subjective assessments showing a more pronounced response and the two systems demonstrating a non-independent relationship.

Methodology

simulator

Sample size: 20

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• passenger effects
• behavioral adaptation risk compensation