A quick check of my calendar shows this is August of 2016. Is my calendar wrong? Well, not to worry. I put on my naysayer hat and discovered the “facts”.
This will never work because Uber did not address snow, fog, heavy rain, dust, pigeons, propane fireballs, or 80-Year-Olds on Roller Skates suddenly veering into traffic. Besides, the car doesn’t have the morals it needs.
Also note the ride will be supervised by humans in the driver’s seat for the “time being.” Of course, naysayers suggest the words “for the time being” mean forever.
Proven Facts vs. Facts of the Matter
For icing on the “this will never work” cake, one reader continually points out “NONE of the manufacturers have any idea how to drive in the snow. In fact, few if any have attempted it. It’s a proven fact“.
Facts of the Matter
Near the end of 2014, Uber co-founder and Chief Executive Officer Travis Kalanick flew to Pittsburgh on a mission: to hire dozens of the world’s experts in autonomous vehicles.
“Travis had an idea that he wanted to do self-driving,” says John Bares, who had run CMU’s National Robotics Engineering Center for 13 years before founding Carnegie Robotics, a Pittsburgh-based company that makes components for self-driving industrial robots used in mining, farming, and the military. “I turned him down three times. But the case was pretty compelling.” Bares joined Uber in January 2015 and by early 2016 had recruited hundreds of engineers, robotics experts, and even a few car mechanics to join the venture. The goal: to replace Uber’s more than 1 million human drivers with robot drivers—as quickly as possible.
And according to most analysts, true self-driving cars are years or decades away. Kalanick begs to differ. “We are going commercial,” he says in an interview with Bloomberg Businessweek.
Starting later this month, Uber will allow customers in downtown Pittsburgh to summon self-driving cars from their phones, crossing an important milestone that no automotive or technology company has yet achieved. Google, widely regarded as the leader in the field, has been testing its fleet for several years, and Tesla Motors offers Autopilot, essentially a souped-up cruise control that drives the car on the highway. Earlier this week, Ford announced plans for an autonomous ride-sharing service. But none of these companies has yet brought a self-driving car-sharing service to market.
Uber’s Pittsburgh fleet, which will be supervised by humans in the driver’s seat for the time being, consists of specially modified Volvo XC90 sport-utility vehicles outfitted with dozens of sensors that use cameras, lasers, radar, and GPS receivers. Volvo Cars has so far delivered a handful of vehicles out of a total of 100 due by the end of the year. The two companies signed a pact earlier this year to spend $300 million to develop a fully autonomous car that will be ready for the road by 2021.
The Volvo deal isn’t exclusive; Uber plans to partner with other automakers as it races to recruit more engineers. In July the company reached an agreement to buy Otto, a 91-employee driverless truck startup that was founded earlier this year and includes engineers from a number of high-profile tech companies attempting to bring driverless cars to market, including Google, Apple, and Tesla. Uber declined to disclose the terms of the arrangement, but a person familiar with the deal says that if targets are met, it would be worth 1 percent of Uber’s most recent valuation. That would imply a price of about $680 million. Otto’s current employees will also collectively receive 20 percent of any profits Uber earns from building an autonomous trucking business.
Otto has developed a kit that allows big-rig trucks to steer themselves on highways, in theory freeing up the driver to nap in the back of the cabin. The system is being tested on highways around San Francisco. Aspects of the technology will be incorporated into Uber’s robot livery cabs and will be used to start an Uber-like service for long-haul trucking in the U.S., building on the intracity delivery services, like Uber Eats, that the company already offers.
Levandowski, one of the original engineers on the self-driving team at Google, started Otto with Lior Ron, who served as the head of product for Google Maps for five years; Claire Delaunay, a Google robotics lead; and Don Burnette, another veteran Google engineer. Google suffered another departure earlier this month when Urmson announced that he, too, was leaving.
The Otto deal is a coup for Uber in its simmering battle with Google, which has been plotting its own ride-sharing service using self-driving cars. Otto’s founders were key members of Google’s operation who decamped in January, because, according to Otto co-founder Anthony Levandowski, “We were really excited about building something that could be launched early.”
Otto Take III
I discussed Otto twice earlier this year. The first time I did not even know I was talking about Otto.
On February 16, I reported that a reader saw a Nevada truck that was completely driverless. I believed the reader.
Comments from skeptical readers were amusing:
TZ said: “Snap a picture next time. Bat boy and the area 51 aliens aren’t very tall, so maybe one of them was driving – apparently aliens can get CDLs.”
Nonetheless, several readers said “Otto is vaporware”.
What about the proven fact: “NONE of the manufacturers have any idea how to drive in the snow. In fact, few if any have attempted it. It’s a proven fact“?
The “fact provider” insisted he was right. That Ford only tested in the rain and trivial amounts of snow. He insisted it was a “proven fact” that driverless cars do not work in the snow and no one knows how to solve the problem.
“Proven Facts”, Please Meet Reality
During a mid-March snowstorm, researchers from MIT Lincoln Laboratory achieved real-time, nighttime, centimeter-level vehicle localization while driving the test vehicle at highway speeds over roads whose lane markings were hidden by the snow. The sport utility vehicle used in the demonstration was equipped with a system that employs a novel ground-penetrating radar technique to localize the vehicle to within centimeters of the boundaries of its lane. The technique could solve one of the problems limiting the development and adoption of self-driving vehicles: how can a vehicle navigate to stay within its lane when bad weather obscures road markings?
“Most self-localizing vehicles use optical systems to determine their position,” explained Byron Stanley, the lead researcher on Lincoln Laboratory’s Localizing Ground-Penetrating Radar (LGPR) program. “They rely on optics to ‘see’ lane markings, road surface maps, and surrounding infrastructure to orient themselves. Optical systems work well in fair weather conditions, but it is challenging, even impossible, for them to work when snow covers the markings and surfaces or precipitation obscures points of reference.”
The Laboratory has developed a sensor that uses very high frequency (VHF) radar reflections of underground features to generate a baseline map of a road’s subsurface. This map is generated during an LGPR-equipped vehicle’s drive along a roadway and becomes the reference for future travels over that stretch of road. On a revisit, the LGPR mounted beneath the vehicle measures the current reflections of the road’s subsurface features, and its algorithm estimates the vehicle’s location by comparing those current GPR readings to the baseline map stored in the system’s memory. An article titled “Localizing Ground Penetrating RADAR: A Step Toward Robust Autonomous Ground Vehicle Localization” in the January issue of the Journal of Field Robotics describes Lincoln Laboratory’s demonstration of the use of LGPR to achieve 4 cm (~1.6 inch) in-lane localization accuracy at speeds up to 60 mph.
“The LGPR uses relatively deep subsurface features as points of reference because these features are inherently stable and less susceptible to the dynamics of the world above,” said Stanley. Surface characteristics, such as lane striping, signs, trees, or the landscape, are dynamic; they are changed or change over time. The natural inhomogeneity in the highly static subterranean geology—for example, differences in soil layers or rock beds—dominates the GPR reflection profiles. The GPR-produced map of the subsurface environment is a fairly complete picture, capturing every distinct object and soil feature that is not significantly smaller than a quarter of a wavelength.
The researchers who conducted the March test runs—Stanley, Jeffrey Koechling, and Henry Wegiel—alternated driving over the route in the Boston area during the snowstorm from midnight to 9:00 a.m. Correlation and overlap estimates demonstrated that the LGPR technique could accurately, and repeatedly, determine the vehicle’s position as it traveled. Post-processing methods on a sensor suite were then used to provide estimates of the accuracy of the system. “We thought the radar would see through snow, but it was wonderful to finally get the data proving it,” said Koechling.
The reliance on static underground features is LGPR’s advantage as a complement to other localization methods, even in fair weather conditions. The use of a subsurface map reduces the need for continual modifications to high-resolution road maps. Fusing GPS, lidar, camera, and LGPR results yields a system that can accurately localize even when one of the sensing modes fails. This “fail-safe” capability will be necessary to the development of dependable autonomous vehicles that can handle demanding ground environments. Autonomous vehicles are outfitted with a range of sensors, such as cameras, radar, and lidar systems, that collect data to enable necessary capabilities—for example, localizing, judging distances, and detecting obstacles. The inclusion of an LGPR system to the sensor suite increases the ability of the vehicle to reliably know its position even when lane markings are hidden by snow, signs and landmarks have moved, and GPS signals are unavailable.
Currently, Lincoln Laboratory is working to explore all weather, GPS-denied, and mapping capabilities. LGPR maps may be useful in helping federal, state, and local governments know when roads and bridges are in need of maintenance. Uses of LGPR for indoor and underground navigation, such as in mining, also remain to be explored. Because of LGPR’s wide range of uses and demonstrated accurate localization in snowstorms, it’s likely that your self-driving car of the future will be equipped with a ground-penetrating radar.
Damn Those “Proven Facts”
Those who believe cars cannot handle snow, fog, heavy rain, and dust storms just might want to consider the MIT video, also shown below.
What the heck is the matter with MIT. Don’t they know “facts” cannot be unproven?
Ford’s Track Record
One clever reader suggested Ford was more likely to go out of business in 2021 than have an completely driverless car by that date.
Actually, the reader has it backwards, Ford is more likely to go out of business when driverless is fully adopted.
A second unthinking reader cited Ford’s hydrogen vehicle that never got off the ground as evidence this would not work.
Competition Yet Again
Ford’s idea had no competition simply because the idea was bad. Look at the competition now.
Players in the Field
Google, Apple, Otto, Uber, GM, Ford, Volvo, MIT, Mercedes, BMW, Toyota, Tesla, other car manufacturers, and dozens if not hundreds of companies are all on a mission to be first.
This is likely the biggest collection of talent focused on a single key idea in the history of mankind.
For the naysayers to be right, every damn one of those companies have to fail.
How Many Jobs Lost?
My statement that “millions of long haul truck driving jobs will vanish in the 2022-2024 time frame” is likely way off on the low side if one counts Uber, taxi, and chauffeur driven vehicles.
Take a look at Uber’s goal once again: “replace Uber’s more than 1 million human drivers with robot drivers—as quickly as possible.”
That’s just Uber. And those jobs will vanish. All of them. What about Lyft? Taxis?
This is going to happen. The components are all in place. Regulation has five years to catch up, and it will. Competition ensures it.
Mike “Mish” Shedlock