When the latest Silicon Valley company involved in self-driving-truck technology sent a big rig on a successful 120-mile autonomous drive earlier this month, there were no conspicuous sensors affixed to the nondescript vehicle, no police escorts guiding the way, and no grand pronouncements following the journey. There were a half-dozen employees on hand, a few thousand pounds of freight in the rear, and miles of open Florida Turnpike ahead.
“It wasn’t some cool marketing stunt,” says Stefan Seltz-Axmacher, taking a not-so-subtle dig at its high-profile competitor Otto, which produced that highly publicized, highly orchestrated delivery of beer in October. Until now, he has been content to keep his own autonomous-trucking effort, Starsky Robotics, out of the spotlight and under wraps.
Seltz-Axmacher founded the company in September 2015, but he hasn’t publicly said much about its intentions. That changed this week, when he discussed at length his plans to develop self-driving trucks that handle long-haul rides between distribution centers and help curb the death toll associated with the industry.
“
“Right now, we’re incredibly focused on not being a
science experiment.”
– Stefan Seltz-Axmacher, Starsky Robotics
”
Starsky Robotics, based in San Francisco, joins an ever growing number of companies that are trying to bring self-driving trucks to the marketplace. The list includes Peloton and Embark as well as Otto, which Uber acquired for $680 million last August. In that crowding field, Starsky may differentiate itself by eschewing the use of lidar in its suite of sensors that help vehicles determine their paths along roads.
“If lidar becomes magical and readily available tomorrow, we will absolutely rethink that decision,” Seltz-Axmacher told Car and Driver. “Right now, we’re incredibly focused on not being a science experiment. It’s not as intensive as Otto, covered in $400 billion worth of lidar, and we’re not a good commercial for [lidar maker] Velodyne. But we could bring this product to the market.”
Using front- and side-facing radar and camera systems for making sense of surroundings in Freightliner Cascadia models, Starsky says it could be ready to deploy commercially by the end of 2017 along specific routes. That time frame may sound compressed, but the company already has been paid for its autonomous-trucking services. In August 2016, one of its trucks moved trailers around a private truck yard in Georgia under autonomous guidance.
Aside from mining vehicles, Seltz-Axmacher believes that job makes Starsky the first company that has done driverless work for money.
While many OEMs and tech companies alike are seeking partnerships with others to help bring autonomous technology to the market, Seltz-Axmacher said that is unlikely.
“I’ve never been a particularly popular person, and I don’t think that will change, so we’re building everything with the intent of not wanting to need many friends,” he said. “We’re not relying on big partnerships or OEMs or magical technology that someone else will finish and build. We’re building everything to not require help, and everything is off the shelf.”
The company has tested on public roads in Florida and has plans for testing in Nevada and Michigan. The test along the Florida Turnpike occurred on February 3, with a freight load of roughly 5000 pounds in the trailer. In that test, the vehicle traveled from the Orlando area along the turnpike toward Fort Lauderdale, with its self-driving system handling driving duties for about 120 miles during the 140-mile trip. As for the remaining 20 miles, the non-autonomous portion of the route fits within Starsky’s intent to focus on long-haul trucking, most likely carrying products between distribution centers.
In that scenario—even as some economists and government officials worry about the wave of unemployment that automated technologies could cause among truck drivers—there may still be an important role for human drivers, who ostensibly would stay employed delivering goods on shorter routes into cities. That’s an ideal scenario, said Seltz-Axmacher, because it keeps drivers employed and closer to their families rather than growing tired on cross-country routes that keep them away from home for weeks at a time. He argues in favor of deploying quickly, using computer vision, radar, and software, because there’s a “moral hazard” in waiting waiting the length of time he believes it would take for lidar to develop.
“One driver told us he’s home only one day a month, and you know, now, you may have drivers who can coach their kids’ soccer games,” he said. “Similarly, there’s a huge safety concern, and truckers are involved in 4000 traffic deaths every year. Thirty percent admit they have fallen asleep behind the wheel. They’re not bad people, but that’s the way the industry is set up. Our drivers will never be faced with that.”
from Car and Driver BlogCar and Driver Blog http://ift.tt/2liuP6C
via IFTTT
0 comments:
Post a Comment