How V2X may have warned Uber of impending crash

Article By : Junko Yoshida

No autonomous car today is designed with the emergence of V2X in mind, but for developers like Uber, V2X might be worth thinking about.

While working on a story about the aftermath of a car crash that involved a driverless Uber vehicle, I came across two automotive industry analysts who brought up the same question: Could V2X have helped the Uber vehicle—which was in driverless mode when it hit a van and rolled over—avoid the collision?

The vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication technologies are collectively known as V2X.

After my initial interview, the Linley Group’s senior analyst, Mike Demler, sent me this message: “There is one thing I forgot to mention. If Tempe was a Smart City with V2X infrastructure and the Ford and (Uber) Volvo had V2X, the DSRC messaging could have warned the Volvo of an impending collision.”

Ian Riches, director of global automotive practice at Strategy Analytics, also wondered whether this outcome would have been different “if both vehicles had some form of V2V technology.”

Uber isn’t offering any insights as to what went wrong in the crash.

Uber accident (cr)
Figure 1: Uber accident scene in Tempe, Arizona (Source: Mark Beach).

However, the company, after suspending its development operations and passenger pilots for the post-crash weekend, put “dozens of” their driverless robo-Ubers on the road in Tempe and Pittsburgh Monday afternoon and resumed passenger pilots. In San Francisco, Uber released two “development vehicles” Monday morning.

We know that no autonomous car today is designed with the emergence of V2X in mind.

But for developers like Uber, V2X might be worth thinking about, as added sensory data input that can make their vehicles safer. The sensors already in driverless cars are there to detect visible objects in the present. V2X can predict the future—allowing the car to sense possible trouble beyond its line of sight, like Peter Parker’s spider-sense.

An alert 3s before the crash?

I called Savari, which makes V2X software. Savari confirmed that the University of Arizona has an active V2X pilot in the Phoenix/Tempe area.

Asked about the Uber vehicle mishap in Tempe, Franz Tschimben, Savari’s strategy honcho, said, "We can’t be 100% sure the crash could have been avoided if both cars had V2X technology."

However, he stressed, “We do know that the driver/tester in the Uber car would have received an alert about three seconds before the crash to take control of the vehicle.”

This would occur because the cars are talking to each other, sending information about their speed and position, and Savari’s software provides “Path Prediction” which would have seen the looming collision if all things stayed the same, he explained.

These three extra seconds turn out significant. Tschimben noted, “Having three extra seconds allows the physical or robot driver to take necessary action to try and avoid the crash. Additionally, the car that caused the accident would have alerted the driver about three seconds before collision to also take action.”

But let’s not forget the obvious.

Even if Tempe was a “smart city” with V2X infrastructure, as Tschimben explained, “You would still need both cars in the described accident above to have V2X technology to try and avoid the crash.”

Therein lies the rub. However, this is possibly a fixable problem, if and when V2X gets mandated.

Just last week, I came across a press release by Autotalks (Kfar Netter, Israel), a developer of V2X chipsets, announcing that the company “raised $30 million (₹203.98 crore) in Round D funding to accelerate global deployment of V2V communication for improving road safety.”

Next: Deadline looms for V2X communication rulemaking »

Leave a comment