For pedestrians, crossing a road often involves reading small signals from an approaching vehicle. A driver may slow down, wave, or make eye contact to show that it is safe to continue.

Automated vehicles do not have a person behind the wheel to provide those familiar cues, creating a new communication challenge. Researchers from Virginia Tech and Zoox are exploring whether specially designed sounds could help fill that gap.

Giving Vehicles a New Voice

The research examined whether sound could tell pedestrians what an automated vehicle is about to do. Instead of relying only on a conventional horn, the team developed seven custom audio tones with different purposes.

Some sounds were intended to communicate that the vehicle was approaching and that a pedestrian should wait. Others were designed to indicate that the vehicle had detected the pedestrian and was waiting, allowing the person to continue across the road. The idea is relatively simple: if people can understand a vehicle's intention through sound, they may not need to depend entirely on visual signals.

Testing Realistic Road Situations

Researchers tested the sounds at the Virginia Smart Roads closed-course facility, creating a controlled environment that resembled real traffic conditions. Forty participants took part as pedestrians, experiencing both marked-crosswalk situations and scenarios where they crossed outside a designated crossing area.

The study included 200 scenarios involving crossings outside marked areas and another 160 crosswalk scenarios. Researchers recorded how participants reacted after hearing different tones and also asked them to evaluate the sounds based on characteristics such as urgency and friendliness.

The test vehicles were capable of automated driving and were equipped with external speakers. A concealed driver was used to operate the vehicle consistently during testing, while participants experienced the interaction as though the vehicle had no visible driver.

Strong Results for Warning Sounds

The clearest results appeared when the audio tones were used to discourage pedestrians from crossing in potentially risky situations. All of the warning tones tested were effective at reducing the willingness of participants to cross. According to the researchers, their performance was comparable to a conventional horn, while participants generally viewed the specially designed sounds more favorably.

Reaction times were particularly notable in the scenarios involving crossings outside marked areas. Participants made the decision not to cross faster across the tested sound conditions than they did under the established horn-based comparison. This finding could be useful because communication between automated vehicles and pedestrians becomes especially important when people encounter vehicles away from traditional intersections.

When the Message Is “Go Ahead”

The results were less consistent when sounds were used to tell pedestrians that the vehicle would wait for them. The waiting tones encouraged people to cross in slightly less than half of the tested crosswalk situations. However, the amount of time participants needed to complete the crossing was not significantly different from the baseline condition.

Researchers suggested that people may need repeated exposure before they naturally understand what a particular sound means. It is also possible that pedestrians can obtain enough information from the vehicle's movement without needing an additional audio message. That distinction is important. Creating a sound is one challenge; designing a sound that people immediately understand is another.

Designing Sounds People Understand

The research team faced the challenge of making the tones noticeable without simply creating another harsh warning. The goal was to develop audio signals that could communicate different intentions while remaining easier to accept than a traditional horn. Researchers also considered people with different levels of vision.

The participant group included people who reported normal or corrected vision as well as people with significant visual limitations. This allowed the study to consider whether sound could provide useful information beyond what a person might obtain visually.

The study suggests that automated vehicles may eventually need their own communication language for interacting with people outside the vehicle. That does not mean these particular tones are ready for widespread road use. The researchers emphasized that additional work is still needed before such sounds could become part of everyday automated driving systems.

Still, the research demonstrates an interesting possibility. As vehicles become increasingly automated, communication may no longer depend on a driver's gestures or eye contact. Carefully designed audio could become another way for vehicles to express what they are doing. For pedestrians, that could make interactions with automated vehicles easier to understand. For vehicle designers, it opens another area of development where sound, technology, and human behavior come together.