Hexagon PI positioned for autonomous success
Note: This article was written for Autonomous Vehicle Technology and was originally posted here. For more information about how Marx Layne & Company can elevate your automotive brand, please visit http://automotive.marxlayne.com.
BY KEVIN JOST; Autonomous Vehicle Technology
Companies like Hexagon PI are increasing their presence in the automotive space as the potential for future mobility presents potentially lucrative business opportunities. This inside look at its strategy includes its acquisition of AutonomouStuff, one of the foundational companies in the autonomous vehicle industry.
As the vehicle industry pushes forward with greater autonomy, locating vehicles by deter- mining their “absolute position” on roads is a critical element. Autonomous driving technology requires pinpoint and reliable accuracy everywhere and at all times, so demand continues to increase for precise positioning technology.
One company that has long-term experience in positioning intelligence is Hexagon PI. Other endeavors where its expertise in precise positioning is applied are surveying and mapping, mining, UAVs, precision agriculture, and marine industries. Its Veripos brand is a leader in positioning for vessels in the marine oil and gas industry.
It hopes to build on that expertise to play a critical role in delivering the necessary high-integrity and high-precision positioning capabilities for autonomous production vehicles. The company has actually been selling ground-truth systems, which establish where the car actually is for test purposes, to all the major OEMs and Tier 1s for many years, said Jonathan Auld, Vice President -Engineering & Safety Critical Systems at Hexagon PI.
But those products are not intended or targeted for serial production applications, said the University of Calgary alum and Novatel veteran, who came back to help the company transition its high-precision positioning technology into the automotive application space.
“We’ve been able to achieve centimeter-level accuracy for decades; we’ve been doing it in a number of different applications,” said Auld. “There’s just never been a use case for something more mass market until some of these recent autonomous applications came out.”
Automotive Tier 1s and OEMs want to know where the car is at the lane level—or about 10-20 cm of precision, said Auld, not the 1-2 cm typically bandied about, but much more precise than today’s GPS receivers that have little better than 5m accuracy. GNSS technology provides that precision, but industry is working on “getting the cost model for business to be able to deliver the accuracy they require in the 10-20 cm range and the safety side of it so that they can use it to build up their safety case when they integrate the GNSS with all sorts of other sensors in the car,” said Auld.
Hexagon PI takes positioning beyond meter-level GPS accuracy to the needed centimeter-level through use of a more-precise GNSS (Global Navigation Satellite System) along with other technologies, including INS (Inertial Navigation Systems), corrections services, and complex algorithms.
GNSS offers global coverage and functions in all-weather scenarios to provide precision positioning. INS, a technology based on measuring the relative motion of platforms using accelerometers and gyroscopes and then integrating that information to figure out relative position and relative position change over time, complements GNSS well. “When one doesn’t work, the other one usually does, and vice versa,” said Auld. “When GNSS is running, and as long as it’s tracking satellites, you’re getting a solution. If you lose lock on the satellites or the sky gets covered up with buildings or overpasses, you can bridge those outages by using inertial technology.”
Hexagon’s algorithms take the GNSS and INS measurements and fuse them into a single solution with continuous position velocity and time output. If there is an ex- tended period of lost GNSS, the INS signal will also degrade over time, other sensor signals (such as LiDARs, cameras, and radars) can be used to continue to output a trajectory.
The lure of automotive
Hexagon PI is just one of the companies coming into the automotive space for autonomous vehicle reasons. This trend of positioning companies that have been outside of automotive, from aviation to surveying, trying to come into automotive is reinforced by the moves of Trimble Navigation, a long-time Hexagon competitor. Its highest profile project is the Super Cruise system that General Motors developed for Cadillac.
The promise of much more business is the lure: “If you can get in the game, the order of magnitude of business certainly changes quite a bit,” said Auld. The auto- motive push fits into the growth plans for the company and its corporate vision of assured positioning anywhere. It builds upon the multiple applications that have aspects of autonomy—like farming and mining. The company is helping the automotive push will help move its technology forward and redeploy it back into established and new markets.
However, there are still challenges on the road to a cost-effective positioning solution that provides a precise location 100% of the time.
“The availability of the correction net- works that enable that accuracy has to be extended into as many regions in the world as possible,” said Auld. “So we’re investing heavily in rolling out our own networks across North America and Europe and eventually Asia to enable that.”
The other part of the puzzle concerns the all-important aspect of safety. “No one’s ever done this kind of accuracy in this difficult an environment with the safety aspects on top of it,” said Auld. He is looking to new processors and standards to help customers to have the required “trustworthiness level.”
For some of the safety aspects, Hexagon PI is also taking some of the concepts of aviation GNSS and extending them into automotive. For the past few years the company has been working with Stanford and the Illinois Institute of Technology, both world-renowned schools in the area of GNSS integrity. It has just begun its third year of funded research at those two institutions “to figure out some of these problems that haven’t been solved around how do you build the safety into the system while maintaining that level of accuracy that’s required,” said Auld.
The big catch
Another major step in Hexagon PI’s bigger entrance into the automotive market was its acquisition last year of AutonomouStuff to embed itself more deeply into autonomous vehicle development.
“Combined with Hexagon PI’s leader- ship in high-accuracy, functionally safe, and high-integrity positioning technology, the addition of AutonomouStuff and their offerings is helping our customers to accelerate the development of more comprehensive Autonomous X solutions,” said Michael Ritter, President and CEO of Hexagon PI. “Our expanded capabilities will allow Hexagon PI to meet the industry’s ever growing demand for more robust autonomy solutions.”
AutonomouStuff, which employs just over 100 people, is headquartered in Morton, IL, with operations in Silicon Valley, Detroit, Chicago, Houston, China, and Europe. Founder and CEO Bobby Hambrick chose Morton because he was born and raised there, his parents both long-time Caterpillar employees.
Hambrick founded AutonomouStuff in 2010 after pitching his idea of bringing his vision of autonomous vehicle development to the company he worked for at the time, which provided factory automation solutions.
“They thought I was crazy and didn’t believe that there was a future in automated driving,” said Hambrick. “I told them that [autonomy was] the beginning of the future of transportation, and that it was going to be reinvented and this is going to be a big opportunity for us to provide services.”
And boy was he right.
“I wanted to pull together products, services, or whatever is required to jumpstart the deployment of automated driving; we call that enabling the future of mobility or autonomy,” he said. “The vision from the very beginning is still true today, but the business to make that happen has changed and does change almost every single day because of the way the industry is adapting and moving very quickly.”
State of AV development
During the DARPA (Defense Advanced Re- search Projects Agency) Challenge days in the mid- to late-2000s, the industry looked at AVs as a “kind of a fun research project, maybe in the military, but it’s too expensive,” said Hambrick. “No one thought that it was really possible. That was just in 2007. A few years later, things changed very quickly. [Deployment was expected] in 2050, and then it was 2030, and now its 2020.”
Hambrick says the technology is evolving incredibly quickly, from the ability to process ever-larger amounts of data to the software infrastructure that’s available to the open-source community.
“Everything is changing faster than anyone ever imagined,” said Hambrick. “In fact, every single company in the world right now is struggling with the pace of innovation. It’s not just specific to what we’re doing with automated driving. That’s the result of this fourth industrial revolution that everyone talks about.”
He singled out a few AV happenings that are changing the way the whole industry is evolving, lowering the barrier of entry, and speeding up the whole process.
He likens the current state of the AV industry to when Google’s Android was re- leased in 2010 and the cell phone and mobile business really started taking off. And he views Baidu’s Apollo as the Android of automated driving, “an independent open- source stack of software that you can put on basically any hardware as long as you know what you’re doing. So that democratized a lot of the complexity and the magic behind automated driving.”
He believes the Uber incident last year in Arizona slowed the innovation down a bit and made people start thinking about functional safety, “but that’s not going to stop them from moving forward and still continuing to invest very heavily.”
What Waymo is doing in mobility services is a prime example: “They’re a good five or so years ahead of everyone else,” he said.
On mobility services in general, “it’ll be a long time before those guys make any money doing that,” but he said “they’re investing heavily for the future.”
He is somewhat more bullish on the short-term prospects for automation in long-haul trucking.
“There’s an obvious business case there,” as evidenced by the number of startup and traditional trucking companies working on automated solutions. “It’s always the last mile that’s the complicated part. You can drive down a highway across Iowa and Nebraska no problem with today’s technology. What do you do when you get there and you’ve got to go on a ramp, pull into a depot, or get gas?”
He believes tele-operation might be an answer, with a number of startup companies looking into OnStar-like solutions.
Synergies and scale
Acquiring AutonomouStuff was an easy way for Hexagon PI to jump deeply into the automated driving space, but it also provided the former with a way to quickly grow.
“What I needed to do to scale and grow, Hexagon had to offer. There’s a big infra- structure worldwide; they’re in 50-some- thing countries.”
One of those countries is China, where AutonomouStuff recently announced a historic partnership with Great Wall Motor Co., China’s largest SUV and pickup manufacturer. The collaboration consists of joint engineering efforts to develop vehicles used for automated driving research and development, and to enable safe, reliable, and secure interfaces to control premium Wey and Haval vehicles.
“Our agreement with Great Wall Motor Company will enable lower development costs and increase safety for automated driving research, thereby significantly reducing the barrier of entry,” said Hambrick.
Despite the Great Wall success, the Chinese market presents a huge challenge for the relatively small AutonomouStuff.
“I was running into huge roadblocks of scaling in China without having infrastructure there. Hexagon has infrastructure there. To take that next step, it is going to take work and investment and infrastructure. So that’s another really good reason why it makes sense for Hexagon and AS to work together.”