By Jeff Sanford
Toronto, Ontario -- August 13, 2017 -- In this week's Autonomous Report, we look into the unexpected high costs of autonomous vehicles (AVs), the deleterious effect of automation on driving skills, a self-driving “hoax” turns out to be a research project in disguise and much, much more!
- One reason not to worry about a sudden flood of AVs has to do with that most basic economic function, price. Austin Russell, CEO of Silicon Valley startup Luminar, recently said in an interview he estimates that the first fully autonomous vehicles will be radically expensive—up to $300,000 or $400,000 per car. According to Russell, “... the technology hasn't progressed as far as some analysts claim.” LiDAR systems are what allow AVs to see around them. According to the story current models are not yet reliable enough.
"People think that they'll go and buy autonomous cars. That's not going to reflect reality,” he was quoted as saying. The story goes on to note that current LiDAR systems have a critical failure rate of about one error in 1,000 miles. “To be acceptable for use on public roads, that rate will have to drop to one in 5 million miles,” according to Russell. So far the trend in the industry has been to focus on bringing down the cost of the devices rather than improving performance. Once companies begin optimizing the technology to necessary levels costs will rise, making fully-autonomous cars “too expensive for most people to own,” according to the story.
- The major financial news publisher Bloomberg wonders if the new flood of assisted driving features are making people worse drivers. According to the Bloomberg story, “Increasingly, automakers are worrying it may.” Today of course driver-assist technology is, “... rapidly spreading from luxury cars to everyday Hondas, Nissans and Chevys.” But according to the story, “... these automated aids aimed at improving safety are having an unintended consequence: They're degrading driving skills.”
Adrian Lund of the Insurance Institute for Highway Safety (IIHS) is quoted as saying, “There are lots of concerns about people checking out and we are trying to monitor that now. Everything we do that makes the driving task a little easier means that people are going to pay a little bit less attention when they're driving." According to the Bloomberg story, “The auto industry is 'terrified' about the unwanted side effects of their popular new features, and companies are scrambling to find ways to keep drivers engaged rather than glued to their smartphones.” GM is, “... installing eye-tracking technology on the Super Cruise feature coming to Cadillac models later this year, which allows drivers to take their hands off the wheel but requires watching the road,” according to the story. As well, “Nissan Motor Co.'s ProPilot Assist keeps the car centered and brings it to a stop in its lane if the driver goes more than 30 seconds without grabbing the wheel. Tesla Inc. last year implemented limits on drivers' ability to go hands-free while using the company's Autopilot system.”
- An early debate in the AV era (way back in 2016) was whether cars could be programmed to carry out the ethically-based decision making of the sort that human drivers do. Some critics said cars would be dangerous road-based missiles without an ethical sense to, say, avoid hitting a bus full of schoolchildren by, instead, driving into a wall. Turns out, there's no need to worry. A new study from The Institute of Cognitive Science at the University of Osnabrückhas finds that, “... the moral decisions humans make while driving are not as complex or context dependent as previously thought ... these decisions follow a fairly simple value-of-life-based model, which means programming autonomous vehicles to make ethical decisions should be relatively easy,” according to a report on Futurism. The story goes on to say that, “The previous assumption was that these types of moral decisions were highly contextual and therefore beyond computational modelling.” But one of the scientists involved in the study was quoted as saying, “... we found quite the opposite. Human behaviour in dilemma situations can be modeled by a rather simple value-of-life-based model that is attributed by the participant to every human, animal, or inanimate object."
- Another report this week finds that a, “... live experiment recently confirmed for the first time what... researchers had predicted based on traffic models and computer simulations: The presence of just a few properly controlled autonomous vehicles can eliminate stop-and-go traffic jams caused by human drivers--along with the accident risks, fuel inefficiencies and increased pollution that such congestion causes.” A team of researchers from Temple University found that, “... self-driving cars and related technology may be even closer to revolutionizing traffic control than previously thought.” One of the researchers was quoted as saying, "Our simulations indicated that it would be possible to eliminate phantom traffic jams if just one in every 20-25 vehicles were self-driving. The instabilities that human drivers unintentionally introduce by their driving behaviour can be cancelled by a single well-controlled autonomous vehicle.”
- An article from Reuters notes that the strategy on AVs among some major automakers is shifting as the big OEMs begin to worry about making a profit off the billions now being invested into the AV sector. According to the report, OEMs are now increasingly realizing they have to work together to fund these billion-dollar programs, rather than compete company-to-company in a bid to control the tech entirely within one organization. According to the report, “BMW and Daimler have announced alliances with suppliers, talking up the virtues of having a bigger pool of engineers to develop a self-driving car. But another motive behind these deals, executives and industry experts [say] ... is a concern that robocars may not live up to the profit expectations that drove an initial investment rush.Carmakers are increasingly looking to forgo outright ownership of future autonomous driving systems in flavor of spreading the investment burden and risk. The trend represents a clear shift in strategy from little more than a year ago when most automakers were pursuing standalone strategies focused on tackling the engineering challenge of developing a self-driving car, rather than on the business case.”
That is, the OEMs are already starting to worry about how they are going to profit off the billions being spent on this tech. The report notes that, “Dozens of companies - including carmakers and tech firms like Google and Uber - are vying for a market which, according to consulting firm Frost & Sullivan, will only make up about 10 to 15 percent of vehicles in Europe by 2030. There are sure to be losers.”
- Several stories appeared in the mainstream media last week about an odd event that occurred in Arlington, Virginia last week. A driver took a look inside what, “... he thought was a driverless van... [and] saw something that terrified him -- human hands and legs protruding from the driver-side seat.” The snooper happened to be a reporter for the local NBC affiliate. He ended up being interviewed on CBC's As It Happens show. "It was like a horror movie because it's so not what you were expecting," said the reporter to the Canadian radio show host. "I couldn't believe what I was actually seeing. I could not believe that there was a half-man, half-car seat that was pretending to be an autonomous vehicle driving around the streets of Arlington, Virginia.”
After the news broke, NBC received a call from the Virginia Tech Transportation Institute (VITI) explaining that the man dressed as a car seat was in costume and working on a research project that was, "studying human behaviour in the presence of new technology in the real world." That is, the school was wondering how other drivers would react to a car with no driver. “Weirded out and ready to talk about it on Canadian radio,” seems to be the answer.