Hazard Notifications for Cyclists: Comparison of Awareness Message Modalities in a Mixed Reality Study

Document Title

Hazard Notifications for Cyclists: Comparison of Awareness Message Modalities in a Mixed Reality Study

Document Information

• Subject Area: Bicycle safety, mixed reality technology, traffic assistance systems
• Keywords: accident prevention, safety notifications, cyclist safety, connected traffic, head-mounted display, vulnerable road users

Research Background and Problem

• Problem and Challenges: Cyclists frequently encounter the danger of car doors opening suddenly ("dooring" accidents), which can lead to severe injuries. While some in-car warning systems and infrastructure modifications exist, they are not always feasible in urban settings, leaving cyclists passively exposed to risks.
• Significance: Studies indicate that this type of accident remains prevalent in Germany and other regions. Notifying cyclists of potential dangers can enhance their safety and help prevent accidents.
• Research Motivation: With the growing adoption of intelligent and connected traffic technologies (e.g., Car2X communication), there is potential to develop wearable support systems for cyclists to ensure safety.
• Related Work: Previous research has focused on intelligent warning systems for car drivers and pedestrians, with limited studies addressing hazard notifications for cyclists. Some initial studies using virtual reality bicycle simulators have explored the feasibility of such warning systems.

Solution

• Method or Solution: The authors developed and tested three types of cyclist warning systems:
  1. Visual-only messages (V): A series of visual symbols and text prompts displayed on a head-mounted display (C-HMD).
  2. Visual messages with auditory alerts (V+AI): Adding beeping sounds to enhance the perception of visual prompts.
  3. Visual messages with voice prompts (V+VM): Adding voice prompts to provide more direct feedback.
• Innovations:
  • Proposed a multimodal (visual + auditory) warning system to enhance cyclists' awareness of potential hazards.
  • Explored the feasibility of using a mixed reality (MR) bicycle simulator for innovative development.
• Implementation Steps:
  • Tested the system with 24 participants in a fixed bicycle simulator and mixed reality laboratory environment.
  • Each participant simulated cycling in various road environments and received different types of hazard notifications based on pre-set vehicle and door scenarios.
  • Delivered continuous warnings at intervals of 9 seconds, 6 seconds, and 3 seconds, followed by a completion notification after passing the hazard.
  • Collected feedback on user experience (UEQ), intuitiveness (INTUI), driving presence, and system preferences.

Research Findings

• Specific Findings:
  • Participants generally preferred the multimodal systems, "visual + auditory alerts (V+AI)" and "visual + voice prompts (V+VM)," over the visual-only system (V).
  • Visual prompts combined with auditory signals (especially beeping sounds, V+AI) were perceived as more effective and less likely to be ignored.
  • Users rated the overall experience and intuitiveness of the system (e.g., ease of use, comfort) highly, particularly for systems with auditory prompts.
• Advantages Compared to Existing Solutions:
  • The system leverages in-vehicle sensors and connected traffic technologies (e.g., Car2X) to provide real-time, personalized feedback.
  • The head-mounted display minimizes interference with cyclists' environmental awareness, improving safety.
• Experimental or Evaluation Results:
  • Simulator evaluation results showed that after receiving hazard warnings, participants significantly increased their lateral distance from parked vehicles, reducing accident risks.
  • The simulated cycling environment was deemed sufficiently realistic, though participants noted the need for validation in real-world road conditions.
• Limitations and Future Directions:
  • The sample size was small (N=24) and did not include a broader age range.
  • The simulator was static and did not incorporate dynamic traffic scenarios, leaving some interaction complexities unexplored.
  • Future research should address the integration of multiple hazard notification priorities, as well as the system's long-term acceptance and trustworthiness.
  • Conducting integration tests under real-world road conditions will be a critical next step in the research.