It was like watching an autonomous vehicle at its best. A Ford Transit passenger van with an empty driver’s seat cruised through heavy traffic in the crowded streets of Arlington, Virginia, without a hitch.
The steering wheel seemed to turn by itself; the vehicle obeyed every traffic rule, signaled for turns, and gave way to other traffic. The most remarkable thing about the event is that it was so unremarkable.
It was unremarkable, that is, except for the local NBC television station which breathlessly reported on every move of this van that seemed to be without a driver. Finally the reporter caught up with the van at a light, only to discover that it wasn’t autonomous at all. Inside the van was the driver disguised as the driver’s seat. You could tell close-up because you could see the driver’s hands and legs.
But from any kind of a distance, the van looked as if it was driving itself. And that was the idea.
What was going on here was a project being conducted by the Virginia Tech Transportation Institute. The goal of the project, which is still ongoing, is to determine the effect of a driverless vehicle on the public to determine if special signs or signals might be required.
“This study is investigating the potential need for additional exterior signals on automated vehicles,” the institute said in its press materials. “This research is relevant for ensuring pedestrians, cyclists, and other drivers are accommodated.”
The research was being conducted with a vehicle that appeared to be driverless, but which actually had a driver for safety reasons. “The driver’s seating area is configured to make the driver less visible within the vehicle, while still allowing him or her the ability to safely monitor and respond to surroundings,” the statement from Virginia Tech explained.
“Development of the test vehicle focused on ensuring driver safety and included several months of piloting and testing the vehicle, first in controlled areas, then in low-density areas and finally in an urban area.”
The vehicle included video cameras trained on the route the vehicle was taking as well as on the exterior so that researchers could watch the reactions of pedestrians, other drivers and cyclists during the apparently autonomous drives. At this point, Virginia Tech isn’t making those videos available to anyone but researchers. However, VTTI said that the results of the research will be made public once it’s finished.
While we weren’t able to see the video from the research vehicle, the news coverage video showed little if any reaction beyond that of the television station’s reporter.
Other drivers seemed to be taking the empty van in stride. Of course this may be due partly to the inclement weather during that part of the test, or it may be due to the fact that traffic in the Washington, DC area frequently moves so slowly that the streets more closely resemble parking lots than active thoroughfares.
The fact that the study is continuing suggests that the streets of Arlington and likely other cities in Virginia have not seen the last of the mystery van and its disguised driver. Perhaps the Virginia Tech team will take its research to locations where traffic actually moves and see if it gets a reaction.
While the Virginia Tech test may seem a little odd at first, it appears to be an integral part of the university’s deep involvement in transportation and the role of autonomous vehicles. Virginia has already invested in smart highways designed to make the use of autonomous vehicles safer and more efficient. It also has a series of test tracks and facilities designed for autonomous vehicle development.
According to the Institute, “If designed well, automated vehicles have considerable potential for reducing congestion, increasing safety, and providing new transportation solutions for people who currently cannot drive.” The findings will be used to help determine the best designs for autonomous vehicles, including the need to notify others.
“This study is one of many being conducted to determine how best to design automated vehicles,” the Institute said in its statement. The statement said that it’s relevant to ensuring that pedestrians, cyclists and other drivers are accommodated.
The idea of testing the public’s reaction to self-driving vehicles may appear slightly amusing at first look, but in reality it’s important. Visual cues are one important input for human drivers and frequently those cues come from other drivers. You can, for example, see the other driver and tell if that driver is looking at you, or whether they seem distracted. Knowing that can help drivers anticipate other drivers’ reactions to safely move through traffic.
But with an autonomous vehicle, there’s no other driver from whom to receive visual cues. Signs or signals on the exterior of the self-driving vehicle will prepare human drivers for the fact that they won’t see those cues. In a sense, it’s akin to the efforts in some communities to add beepers or other noise makers to electric cars to alert visually impaired pedestrians to their presence.
At this point we don’t know what Virginia Tech is learning with this research, but finding ways to ensure that other drivers or pedestrians are included is important in itself if autonomous cars are to provide the traffic safety improvements that are being promised.