Not all explicit cues help communicate: Pedestrians' perceptions, fixations, and decisions toward automated vehicles with varied appearance

Overview

• This study investigates how the design and behavior of autonomous vehicles (AVs) influence pedestrians' road-crossing decisions and behaviors.
• The researchers conducted a video-based eye-tracking experiment with 61 participants, who responded to various simulated AV scenarios on an unsignalized, two-way road.
• The study manipulated the AV's kinematic patterns (yielding or non-yielding) and its exterior design (conventional vehicle, AV with no driver, AV with text-based identity, AV with roof sensors, AV with dynamic external human-machine interfaces [eHMIs]).
• The researchers measured participants' perceived clarity, crossing initiation distance, crossing decision time, and gaze behavior during the interactions.

Plain English Explanation

When it comes to road safety, pedestrians are quite vulnerable. So, this study wanted to understand how the look and behavior of self-driving cars (or AVs) might impact how pedestrians decide to cross the road.

The researchers showed people a series of videos depicting different types of vehicles approaching a spot where someone could cross the street. Sometimes the vehicle would yield (stop) to let the pedestrian cross, and sometimes it wouldn't. The vehicles also had different appearances - some looked like regular cars with drivers, others looked like self-driving cars with no driver, and some had special features like text or sensors on the roof.

As the people watched the videos, the researchers tracked their eye movements and recorded how quickly they decided to cross the road, how far away they started crossing, and how confident they felt about the situation. This allowed the researchers to see how the vehicle's behavior and design affected the pedestrians' decisions and actions.

The key finding was that the AV's driving behavior - whether it yielded or not - was the most important factor in how pedestrians decided to cross. But the design of the AV also mattered. Having clear, dynamic displays that showed the vehicle was aware of the pedestrian and intending to stop made the crossing process smoother and more efficient. On the other hand, some design elements like text labels or roof sensors actually distracted pedestrians and made the crossing more difficult.

Overall, this research provides important insights into how self-driving car design can be optimized to improve safety and communication between AVs and pedestrians, especially in complex urban environments. By understanding pedestrian behavior, designers can create AV features that make the roads safer for everyone.

Technical Explanation

The researchers conducted a video-based eye-tracking study to investigate the influence of AV exterior design, correlated with their kinematics, on pedestrians' road-crossing perception and decision-making.

The experiment involved 61 participants who responded to video stimuli depicting a manipulated vehicle approaching a predefined road-crossing location on an unsignalized, two-way road. The vehicle's kinematic pattern was either yielding or non-yielding, and its external appearance was varied across five types: conventional vehicle with a human driver, AV with no driver, AV with text-based identity indications, AV with roof radar sensors, and AV with dynamic eHMIs adjusted to vehicle kinematics.

During the interactions, the researchers recorded participants' perceived clarity, crossing initiation distance (CID), crossing decision time (CDT), and gaze behavior. The results showed that the AV's kinematic profiles played a dominant role in pedestrians' road-crossing decisions, as evidenced by their subjective evaluations, CID, CDT, and gaze patterns.

Furthermore, the use of clear eHMIs, such as dynamic pedestrian icons, reduced pedestrians' visual load, enhanced their perceptual clarity, expedited road-crossing decisions, and improved overall crossing efficiency. In contrast, textual identity indications and roof radar sensors had no significant effect on pedestrians' decisions but negatively impacted their visual attention, as indicated by heightened fixation counts and prolonged fixation durations, particularly under yielding conditions.

Critical Analysis

The study provides valuable insights into how AV design and behavior can influence pedestrian decision-making and crossing behavior. However, there are a few potential limitations and areas for further research:

• The experiment was conducted in a simulated environment, which may not fully capture the complexity and unpredictability of real-world road crossing scenarios. Further research is needed to validate the findings in actual field studies.

• The study focused on a specific road configuration (unsignalized, two-way) and did not consider other common road layouts, such as intersections or multilane roads. Examining pedestrian responses in a variety of road environments would provide a more comprehensive understanding.

• The study did not investigate the long-term effects of familiarity with AV designs and behaviors. As pedestrians become more accustomed to interacting with AVs, their perceptions and decision-making processes may evolve, which could impact the generalizability of the findings.

• The researchers acknowledged that the use of textual identity indications and roof-mounted sensors may have negative effects on pedestrians' visual attention, but did not explore potential design solutions to mitigate these issues. Investigating alternative ways to convey AV identity and sensor information could lead to more effective communication strategies.

Despite these limitations, this study represents an important step in understanding the complex dynamics between AVs and vulnerable road users, such as pedestrians. The findings highlight the crucial role of AV design and communication in shaping pedestrian behavior and can inform the development of safer and more user-friendly autonomous vehicle systems.

Conclusion

This study provides valuable insights into how the design and behavior of autonomous vehicles (AVs) can influence pedestrians' road-crossing decisions and behaviors. The key findings suggest that the AV's kinematic profile (yielding or non-yielding) is the primary factor in pedestrian decision-making, but the vehicle's exterior design also plays a significant role.

Specifically, the use of clear and dynamic external human-machine interfaces (eHMIs), such as pedestrian icons, can enhance pedestrians' perceptual clarity, expedite their crossing decisions, and improve overall crossing efficiency. In contrast, design elements like textual identity indications and roof-mounted sensors, while not directly affecting pedestrian decisions, can negatively impact their visual attention and cognitive resources.

These findings have important implications for the development of safe and user-friendly AV systems, particularly in complex urban environments where pedestrians are vulnerable. By understanding the factors that influence pedestrian behavior, AV designers and policymakers can work to create solutions that prioritize the safety and comfort of all road users. Continued research in this area, including field studies and investigations of long-term familiarity effects, will be crucial to ensuring a smooth and seamless integration of autonomous vehicles into our transportation networks.