Safety Considerations Associated with Driverless Shuttles

Driverless vehicles present an enticing opportunity for increased traffic safety and reduced labor costs. However, technical issues and concerns remain regarding the interaction between driverless vehicles and human drivers. This project included observing and analyzing the interactions of a driverless shuttle and other road users in a pedestrian-heavy area to develop strategies for decreasing potential threats. Increasing the speed of the driverless shuttle and displaying clearer electronic messaging on the back of the shuttle could improve safety challenges.

While driverless vehicles in high-density, walkable neighborhoods offer many potential fiscal and environmental benefits, public buy-in will only occur if these vehicles are widely regarded as a safe transportation alternative. To decrease the potential danger associated with driverless vehicles and increase public perception of them, MnDOT sought to better understand how driverless vehicles interact with human drivers in a pedestrian-heavy area.

“The results of this project provide a clear direction for making the use of future driverless shuttles safer in areas with high pedestrian traffic,” said Cory Johnson, program lead, MnDOT Connected and Automated Vehicles Program.

Two driverless shuttles, known as the Med City Mover, operating in a pedestrian-heavy area of Rochester, Minnesota, provided an opportunity to assess driver behavior near automated vehicles, especially in environments in which a shuttle regularly stops to yield to pedestrians. Anecdotal evidence had indicated that drivers of manual vehicles became impatient and frustrated when following behind the slow-moving shuttles and engaged in dangerous passing that created threats of pedestrian-involved, rear-end and sideswipe crashes.

By better understanding how drivers behave in relation to driverless vehicles and what dangers this behavior creates, MnDOT can develop strategies to mitigate these dangers and help ensure the safety of those in and around driverless vehicles.

What Did We Do?

Preliminary steps included performing a literature review that focused on safety and driver behavior and conducting interviews with driverless shuttle operators and manufacturers. From there, research focusing on shuttle operations included:

Observing driver behavior in a pedestrian-heavy area of Rochester when the driverless shuttle was present and when it was absent.
Assessing driver response and overtaking behavior to varying shuttle speeds at the Minnesota State Fair, and conducting a survey to gauge driver response to varying shuttle signals and messaging, including a turn signal, hazard lights (or emergency flashers) similar to those used by the shuttle in Rochester, Ped X-ing message and Pedestrian Crossing message.
Conducting driving simulations to measure driver response to different types of signals and messaging from the shuttle. This included analyzing driver responses to hazard lights, a turn signal and text/icon messaging on an LED screen on the back of the shuttle.

What Did We Learn?

The research conducted as part of this project provided insight into the safety relationship between driverless shuttles and other road users. Below are highlights from the project findings:

The slower driverless shuttle created a longer queue of vehicles behind it compared to other vehicles along the route.
There was a higher rate of overtaking the driverless shuttle when yielding to pedestrians compared to other vehicles along the route.
Drivers were more likely to pass the driverless shuttle due to both the slower speed and the confusing messaging of the shuttle.
A higher rate of misinterpretation of the shuttle’s behavior and a lower rate of waiting behind the shuttle occurred when using hazard lights (or emergency flashers) compared to using text/icon messaging.
Text/icon messaging resulted in a higher rate of drivers stopping behind a shuttle when yielding to a pedestrian. For example, drivers presented with text/icon messaging were almost 40% less likely to commit a dangerous pass around the shuttle than drivers presented with hazard lights.

What’s Next?

From the results of this project, investigators recommended operating the driverless shuttles at higher speeds and replacing the use of hazard lights (or emergency flashers) with text/icon messaging on the back of the shuttle to better communicate yielding to pedestrians with surrounding vehicles.

While driverless shuttles are not yet in frequent use in Minnesota, this research provides strategies and direction for enhancing safety for future shuttle operations and improving public opinion of driverless vehicles. Further, advances in shuttle technology for speed and messaging will enhance the strategies that resulted from this research.