Thursday, 30 March 2017

Uber Safety Driver: “No Time To React” before Autonomous Crash

Leave a Comment
http://ift.tt/2no6zj2

Mandatory credit: FRESCO NEWS / Mark Beach Uber crash

A driver turns left across multiple lanes at an intersection. Traffic obscures his or her view. An oncoming car approaches unseen. A collision occurs.

That’s a scenario that’s all too common among the estimated 6.3 million traffic crashes that take place each year on U.S. roads, and it’s the same one that resulted in a collision last week that involved one of Uber’s self-driving vehicles in Tempe, Arizona.

More details from the incident emerged Wednesday, when the Tempe Police Department released the full crash report. Consistent with preliminary findings, the report said a human driver was to blame for the four-car crash.

According to police, Alexandra S. Cole drove her 2008 Honda CR-V northbound on South McClintock Drive, a main arterial road in the city. She turned left at the intersection of East Don Carlos Avenue and struck the Uber autonomous vehicle, a specially equipped 2017 Volvo XC90 SUV, which had been traveling through the intersection in the far right of three southbound lanes.

“There was no time to react, as there was a blind spot created by the line of traffic.” – Patrick R. Murphy, Uber

Cole plowed into the left side of the XC90, which subsequently struck a traffic-signal pole and two more vehicles before coming to rest on its right side. She was cited for failure to yield.

“As far as I could tell, the third lane had no one coming in it, so I was clear to make my turn,” Cole wrote in the report. “Right as I got to the middle lane about to cross the third, I saw a car flying through the intersection, but I couldn’t brake fast enough to completely avoid the collision.”

Operating in autonomous mode, the Uber vehicle was traveling at 38 mph, according to Patrick R. Murphy, the Uber safety driver behind the wheel at the time of the crash. The posted speed limit on that area of McClintock Road is 40 mph, according to the police report. From his vantage point, Murphy said, he didn’t see the other car coming until it was too late.

“As I entered the intersection, I saw the vehicle turning left on [Don Carlos Avenue] from the northbound McClintock lane,” he wrote in the police report. “There was no time to react, as there was a blind spot created by the line of traffic in the southbound left lane on McClintock.”

Uber-crash-1

The car was part of a pilot project Uber is conducting in both Tempe and Pittsburgh, Pennsylvania, in which users can summon the autonomous vehicles in the same manner they regularly use the ride-hailing service. No passengers were in the vehicle at the time of the crash.

As far as the humans involved are concerned, the causes of the crash are clear. What remains unanswered is how Uber’s self-driving system sensed and reacted to the sudden threat posed by the turning CR-V. Although the XC90 was found not at fault, the police report offers no details on how the software may have attempted to avoid or mitigate the crash.

An Uber spokesperson said the company fully cooperated with the Tempe Police Department, but remained tight-lipped on whether Uber shared sensor data or camera images with crash investigators. Tempe police did not comment on whether they had received camera or sensor information from the company.

“The interaction between robot cars and human-driven cars is an area of serious concern and requires research.”
– John Simpson, Consumer Watchdog

One of the bedrock promises of self-driving vehicles is that they will reduce the number of overall motor-vehicle crashes and the severity of the ones that cannot be avoided. Executives from any number of autonomous-technology companies regularly cite federal research that suggests 94 percent of traffic crashes are caused by human error or behavior, and they tout safety improvements as a core reason for pursuing self-driving deployments.

Data from the self-driving system involved in the Tempe crash could be a valuable learning tool for both Uber and others, which is why John Simpson, privacy director at Consumer Watchdog, a nonprofit organization that tracks autonomous developments, thinks it’d be a mistake if the company didn’t share the information.

“The interaction between robot cars and human-driven cars is an area of serious concern and requires research,” he said. “That’s why all the data about crashes must be made public.”

In California, regulations mandate that self-driving technology companies file a report with the state’s Department of Motor Vehicles if one of their cars is involved in a crash. Arizona has no such requirement. Arizona’s governor, Doug Ducey, legalized autonomous testing via executive order in August 2015 and has often touted the state’s lack of regulatory requirements in attempts to woo testing from rival California.

Uber paused its testing in the state following the crash and commenced a review of its fleet, but the hiatus was short-lived. The company resumed operations Monday.

Let's block ads! (Why?)



from Car and Driver BlogCar and Driver Blog http://ift.tt/2nDIrce
via IFTTT

0 comments:

Post a Comment