By Jeff Sanford
Toronto, Ontario — February 5, 2017 — This week in the autonomous vehicle (AV) space, we look at the 2,578 failures experienced by AVs in California in 2016, GM shows off its Canadian Technical Centre – Markham Campus to members of the Business Council of Canada, the coming age of autonomous shipping and much, much more!
– Every January the California Department of Motor Vehicles (DMV) releases a major report summing up the performance of AVs. According to a report on IEEE Spectrum, the January report sees, “… carmakers testing self-driving cars in California have to detail how many times their vehicles malfunctioned during the preceding year.” This report tallies up the number of times a self-driving car had to be “disengaged”, that is, “… every time a human safety driver had to quickly take control of their car, either due to hardware or software failure or because the driver suspected a problem.” These disengagement reports indicate how the technology is progressing. This year’s report seems to indicate that Waymo (formerly Google) is pulling ahead of the competition in terms of development of true AV ability. The Waymo cars have performed 635,868 miles of testing accounting, “… for over 95 percent of all miles driven by self-driving cars in California in 2016.” The Waymo fleet of 60 self-driving cars reported, “… a total of 124 disengagements, 51 of them due to software problems.” At the opposite end of the list from Waymo was Bosch, “… which reported over 1,400 disengagements while covering just 983 miles in three vehicles—equivalent to a hefty 1,467 disengagements for every 1,000 miles of driving.” Overall, “a self-driving car failed about every 3 hours in California according to figures …”
The report details a total of 2,578 failures among the nine companies carrying out road-testing in California in 2016. Generally the technology seems to be improving. The number of disengagements in 2016 represents a “sharp reduction” from the number of disengagements in 2015. The stats show a reduction from, “… 0.8 disengagements [in 2015] for every 1,000 miles of autonomous driving down to just 0.2 [in 2016].”
Another company doing well in the tests is Cruise Automation, which is now owned by General Motors. “In its first year on the state’s roads, Cruise’s two dozen cars went from covering less than 5 miles in June 2015 to over 2,000 miles in September 2016. Its disengagement rate also plummeted over the same period, from over 500 to under 3 per 1,000 miles,” according to a report. The same story notes that, “No other company drove more 5,000 miles in 2016, and some of the world’s biggest carmakers, including BMW, Ford, and Mercedes-Benz, covered less than 1,000 … Despite holding testing permits, Honda and Volkswagen drove no autonomous miles at all last year on public roads in California, preferring to test on private courses or out of state.”
– The AV story is part of a larger story about the rapidly evolving technologies around robotics, artificial intelligence (AI) and increasing automation. As sensors are distributed throughout the environment and wired into the network the “Internet of Everything” is coming to be. This is going to be a weird new world for insurance companies, which will be able to measure and detect the “state of the world” at a much finer grain. According to a story on Yahoo! Finance last week, “More than one in five Americans to use activity trackers by 2012.” The story goes on to say that within five years “fitness and health wearables will become a $20 billion market … players who are able to make their devices and systems part of healthcare practices will reap the largest rewards.”
– General Motors Canada held a Business Council of Canada meeting during the launch of the company’s new Canadian Technical Centre – Markham Campus. The campus is where GM will develop “… new autonomous vehicle software and controls, active safety and vehicle dynamics systems, infotainment and connected vehicle technologies.” According to a press release, GM Canada announced in June 2016 that it will, “… increase its eco system of Canadian engineering.” GM Canada President and Managing Director, Steve Carlisle, was quoted as saying, “It is an honour to welcome members of the Business Council of Canada to visit and learn about our new advanced engineering and manufacturing operations in Markham today. At GM Canada we are excited to be giving Canada’s best and brightest software engineers opportunities to help define some of the most important mobility technology changes in a generation.”
– One of the key technologies for the coming era of AVs will be LiDAR, laser-based radar. One of the major makers of the devices is Velodyne. It was reported last week that the company is, “expanding a megafactory in San Jose, Calif.,” as a way of feeding the market this key bit of AV tech. Cash from Ford will help. The expansion of the factory will allow Velodyne to ramp up to 1 million units annually by 2018. LiDAR devices are those spinning things seen on the roof of many delivery trucks today. The systems are expensive (they can cost as much as $73,000). But according to a report the expansion of the factory, “… means that its rooftop tower, the Cadillac of LiDAR, may soon sell at Chevrolet prices.” According to a report, last month Velodyne annouced a “breakthrough” in solid-state design that, “… with mass production, should drop the unit price below $50.” The story goes on to say that, “such cheap-as-dirt LiDAR would find its way into just about every vehicle, bicycles included.”
– George Hotz, a twenty-something hacker, has released a freeware software suite that he says, “… turns a modestly smart car into one that can more or less drive itself. He contends that it outperforms any production car except a Tesla—at a cost, in parts, of just US $700,” according to a report. The story notes that, “Hotz originally planned to sell an aftermarket system for a bit more than that and make a profit. However, that strategy foundered when the National Highway Traffic Safety Agency started breathing down his neck. So he went open source, reasoning that NHTSA can recall only a product that’s sold, not one that’s given away. Anyone can download the code from his startup, comma.ai.”
The system uses radar that is already in the car. According to Hotz the big OEMs think they are going to develop AV software. But Hotz is skeptical. He thinks the OEMs won’t be able to create the software and will end up simply making the cars. That is, rather than offering an operating system to a smartphone like Apple they will simply be akin to the handset makers in Asia. Says Hotz, during meetings, “… with most of the CEOs of the big auto companies…A lot of them are just hopeless, all adamantly convinced they must own self-driving technology, in house. Bill Ford, chairman of Ford—I’m paraphrasing here—said that otherwise they’d all be reduced to being a dumb manufacturer, saying he didn’t want to be just the handset maker. But self-driving cars is a huge computer-science problem, and I just don’t see any part of Ford that’s ever solved a computer science problem before.”
– A report on IEEE Spectrum mentions another new and up-coming tech company, an Israeli start-up Oryx Vision is said to have a technology that, “… splits the difference between radar and lidar. Like a lidar, it uses a laser to illuminate the road ahead, but like a radar it treats the reflected signal as a wave rather than a particle … Today, radars can see to 150- or 200 meters, but they don’t have enough resolution. Lidar provides great resolution but is limited in range to about 60 meters, and to as little as 30 meters in direct sunlight … Oryx’s coherent optical radar to accurately locate debris in the road at 60 meters, pedestrians at 100 meters, and motorcycles at 150 meters—significantly better than the performance of today’s sensor systems.”
– AVs are having trouble detecting bicycles. According to a story in the tech press, “Robotic cars are great at monitoring other cars, and they’re getting better at noticing pedestrians, squirrels, and birds. The main challenge, though, is posed by the lightest, quietest, swerviest vehicles on the road … Bicycles are probably the most difficult detection problem that autonomous vehicle systems face …” A visual computing expert quoted in the story says bikes pose a, “… complex detection problem because they are relatively small, fast and heterogenous … That’s why the detection rate for cars has outstripped that for bicycles in recent years. Most of the improvement has come from techniques whereby systems train themselves by studying thousands of images in which known objects are labeled. One reason for this is that most of the training has concentrated on images featuring cars, with far fewer bikes.”
– A story in the shipping trade media notes that AV ships are likely to be in use long before AV trucks. The writer of the story, Oskar Levander, an executive with major marine engine maker Rolls-Royce, “… anticipates that the first commercial vessel to navigate entirely by itself could be a harbor tug or a ferry designed to carry cars the short distance across the mouth of a river or a fjord and that it or similar ships will be in commercial operation within the next few years. And we expect fully autonomous oceangoing cargo ships to be routinely plying the world’s seas in 10 or 15 years’ time.” The story goes on to say that, “The size of ships’ crews has been decreasing for centuries, and soon some large vessels may sail without anyone on board. Ships designed to be controlled remotely or to navigate themselves will look distinctly different from today’s ships … between 75 and 96 percent of marine accidents are a result of human error, often a result of fatigue. Remotely controlled and autonomous ships would reduce the risk of such mistakes and along with it the risk of injury and even death to crew members, not to mention the dangers to the ship itself … The threat posed by piracy to ships and their crews would also be reduced. That’s because uncrewed ships could be built so that they’d be very difficult to board on the high seas. Even if pirates got aboard, access to the controls could be made unavailable. Indeed, the computers in command could immobilize the ship or have it steam in a circle, making it relatively easy for naval authorities to reach it. Recapture would also be easier than is usually the case in such situations because there would be no crew held hostage. And without a captured crew to ransom, the target of the piracy is significantly less valuable … Expect such a ship by 2020. By 2025, some forward-thinking shipping companies will be operating remotely controlled, completely uncrewed vessels on the high seas. Five years beyond that, uncrewed oceangoing vessels will be commonplace.”
