Wij willen met u aan tafel zitten en in een openhartig gesprek uitvinden welke uitdagingen en vragen er bij u spelen om zo, gezamelijk, tot een beste oplossing te komen.
Oftewel, hoe kan de techniek u ondersteunen in plaats van dat u de techniek moet ondersteunen.
Modern agriculture involves fields of mind-boggling size, and spraying them efficiently is a serious operational challenge. Pyka is taking on the largely human-powered spray business with an autonomous winged craft and, crucially, regulatory approval.
Just as we’ve seen with DroneSeed, this type of flying is risky for pilots, who must fly very close to the ground and other obstacles, yet also highly susceptible to automation; That’s because it involves lots of repetitive flight patterns that must be executed perfectly, over and over.
Pyka’s approach is unlike that of many in the drone industry, which has tended to use multirotor craft for their maneuverability and easy take-off and landing. But those drones can’t carry the weight and volume of pesticides and other chemicals that (unfortunately) need to be deployed at large scales.
The craft Pyka has built is more traditional, resembling a traditional one-seater crop dusting plane but lacking the cockpit. It’s driven by a trio of propellers, and most of the interior is given over to payload (it can carry about 450 pounds) and batteries. Of course, there is also a sensing suite and onboard computer to handle the immediate demands of automated flight.
Pyka can take off or land on a 150-foot stretch of flat land, so you don’t have to worry about setting up a runway and wasting energy getting to the target area. Of course, it’ll eventually need to swap out batteries, which is part of the ground crew’s responsibilities. They’ll also be designing the overall course for the craft, though the actual flight path and moment-to-moment decisions are handled by the flight computer.
Example of a flight path accounting for obstacles without human input
All this means the plane, apparently called the Egret, can spray about a hundred acres per hour, about the same as a helicopter. But the autonomous craft provides improved precision (it flies lower) and safety (no human pulling difficult maneuvers every minute or two).
Perhaps more importantly, the feds don’t mind it. Pyka claims to be the only company in the world with a commercially approved large autonomous electric aircraft. Small ones like drones have been approved left and right, but the Egret is approaching the size of a traditional “small aircraft,” like a Piper Cub.
Of course, that’s just the craft — other regulatory hurdles hinder wide deployment, like communicating with air traffic management and other craft; certification of the craft in other ways; a more robust long-range sense and avoid system and so on. But Pyka’s Egret has already flown thousands of miles at test farms that pay for the privilege. (Pyka declined to comment on its business model, customers or revenues.)
The company’s founding team — Michael Norcia, Chuma Ogunwole, Kyle Moore and Nathan White — comes from a variety of well-known companies working in adjacent spaces: Cora, Kittyhawk, Joby Aviation, Google X, Waymo and Morgan Stanley (that’s the COO).
The $ 11 million seed round was led by Prime Movers Lab, with participation from Y Combinator, Greycroft, Data Collective and Bold Capital Partners.
We’re still very much in the collaboration phase of autonomous driving, since it’s looking still quite a ways off from being anything consumers can use on the regular. That means there’s plenty of opportunity for things like the new “Autonomous Vehicle Computing Consortium” (AVCC) announced today to form. This industry group includes Arm, Bosch, Continental, GM, Toyota, Nvidia, NXP, and Denso, collecting top automakers along with some of the leading chipmakers and tier 1 suppliers in automotive today.
The group’s goal is to work together in order to “solve some of the most significant challenges to deploy self-driving vehicles at scale,” which pretty clearly translates into putting together the collective efforts of some of those who stand to gain most from autonomy becoming a commercially viable technology, in order to speed up said commercialization. While self-driving has been an area of intense investment and focus in the past few years, it still has a ways to go before these companies can start really reaping the rewards of their investments in terms of revenue-driving businesses.
So what will they actually do to achieve this goal? Step one will be setting up a set of recommended specs, essentially, outlining what size, temperature, power consumption and safety standards AV system architectures and computers should adhere to. The idea is that by arriving at some baseline standards, the group will be better able to move from prototyping, which is expensive and low-volume in terms of output, to manufacturing and deploying AVs at the scale where they’ll truly become a viable commercial enterprise.
There’s more to this industry collaboration than just figuring out the specs range for systems, however: Participating companies will “study common technical challenges,” as well, meaning they’ll be putting their heads together to overcome the major, fundamental tech challenges that still act as hurdles to be overcome in getting self-driving vehicles on the roads.
And of course, though the initial founding group includes only those companies listed above, this new group is also open to new members.
Earlier this month, TechCrunch held its inaugural Mobility Sessions event, where leading mobility-focused auto companies, startups, executives and thought leaders joined us to discuss all things autonomous vehicle technology, micromobility and electric vehicles.
Extra Crunch is offering members access to full transcripts of key panels and conversations from the event, such as Megan Rose Dickey‘s chat with Voyage CEO and co-founder Oliver Cameron and Uber’s prediction team lead Clark Haynes on the ethical considerations for autonomous vehicles.
Megan, Oliver and Clark talk through how companies should be thinking about ethics when building out the self-driving ecosystem, while also diving into the technical aspects of actually building an ethical transportation product. The panelists also discuss how their respective organizations handle ethics, representation and access internally, and how their approaches have benefited their offerings.
Clark Haynes: So we as human drivers, we’re naturally what’s called foveate. Our eyes go forward and we have some mirrors that help us get some situational awareness. Self-driving cars don’t have that problem. Self-driving cars are designed with 360-degree sensors. They can see everything around them.
But the interesting problem is not everything around you is important. And so you need to be thinking through what are the things, the people, the actors in the world that you might be interacting with, and then really, really think through possible outcomes there.
I work on the prediction problem of what’s everyone doing? Certainly, you need to know that someone behind you is moving in a certain way in a certain direction. But maybe that thing that you’re not really certain what it is that’s up in front of you, that’s the thing where you need to be rolling out 10, 20 different scenarios of what might happen and make certain that you can kind of hedge your bets against all of those.
For access to the full transcription below and for the opportunity to read through additional event transcripts and recaps, become a member of Extra Crunch. Learn more and try it for free.
Rose Dickey: I’m here with Oliver Cameron of Voyage, a self-driving car company that operates in communities, like retirement communities, for example. And with Clark Haynes of Uber, he’s on the prediction team for autonomous vehicles.
So some of you in the audience may remember, it was last October, MIT came out with something called the moral machine. And it essentially laid out 13 different scenarios involving self-driving cars where essentially someone had to die. It was either the old person or the young person, the black person, or the white person, three people versus one person. I’m sure you guys saw that, too.
So why is that not exactly the right way to be thinking about self-driving cars and ethics?
Haynes: This is the often-overused trolley problem of, “You can only do A or B choose one.” The big thing there is that if you’re actually faced with that as the hardest problem that you’re doing right now, you’ve already failed.
You should have been working harder to make certain you never ended up in a situation where you’re just choosing A or B. You should actually have been, a long time ago, looking at A, B, C, D, E, F, G, and like thinking through all possible outcomes as far as what your self-driving car could do, in low probability outcomes that might be happening.
Rose Dickey: Oliver, I remember actually, it was maybe a few months ago, you tweeted something about the trolley problem and how much you hate it.
Cameron: I think it’s one of those questions that doesn’t have an ideal answer today, because no one’s got self-driving cars deployed to tens of thousands of people experiencing these sorts of issues on the road. If we did an experiment, how many people here have ever faced that conundrum? Where they have to choose between a mother pushing a stroller with a child and a regular, normal person that’s just crossing the road?
Rose Dickey: We could have a quick show of hands. Has anyone been in that situation?
Cao Xudong turned up on the side of the road in jeans and a black T-shirt printed with the word “Momenta,” the name of his startup.
Before founding the company — which last year topped $ 1 billion in valuation to become China’s first autonomous driving “unicorn” — he’d already led an enviable life, but he was convinced that autonomous driving would be the real big thing.
Cao isn’t just going for the moonshot of fully autonomous vehicles, which he says could be 20 years away. Instead, he’s taking a two-legged approach of selling semi-automated software while investing in research for next-gen self-driving tech.
Cao, pronounced ‘tsao’, was pursuing his Ph.D. in engineering mechanics when an opportunity came up to work at Microsoft’s fundamental research arm in Asia, putatively the “West Point” for China’s first generation of artificial intelligence experts. He held out there for more than four years before quitting to put his hands on something more practical: a startup.
“Academic research for AI was getting quite mature at the time,” said now 33-year-old Cao in an interview with TechCrunch, reflecting on his decision to quit Microsoft. “But the industry that puts AI into application had just begun. I believed the industrial wave would be even more extensive and intense than the academic wave that lasted from 2012 to 2015.”
In 2015, Cao joined SenseTime, now the world’s highest-valued AI startup, thanks in part to the lucrative face-recognition technology it sells to the government. During his 17-month stint, Cao built the company’s research division from zero staff into a 100-people strong team.
Before long, Cao found himself craving for a new adventure again. The founder said he doesn’t care about the result as much as the chance to “do something.” That tendency was already evident during his time at the prestigious Tsinghua University, where he was a member of the outdoors club. He wasn’t particularly drawn to hiking, he said, but the opportunity to embrace challenges and be with similarly resilient, daring people was enticing enough.
And if making driverless vehicles would allow him to leave a mark in the world, he’s all in for that.
Make the computer, not the car
Cao walked me up to a car outfitted with the cameras and radars you might spot on an autonomous vehicle, with unseen computer codes installed in the trunk. We hopped in. Our driver picked a route from the high-definition map that Momenta had built, and as soon as we approached the highway, the autonomous mode switched on by itself. The sensors then started feeding real-time data about the surroundings into the map, with which the computer could make decisions on the road.
Momenta staff installing sensors to a testing car. / Photo: Momenta
Momenta won’t make cars or hardware, Cao assured. Rather, it gives cars autonomous features by making their brains, or deep-learning capacities. It’s in effect a so-called Tier 2 supplier, akin to Intel’s Mobileye, that sells to Tier 1 suppliers who actually produce the automotive parts. It also sells directly to original equipment manufacturers (OMEs) that design cars, order parts from suppliers and assemble the final product. Under both circumstances, Momenta works with clients to specify the final piece of software.
Momenta believes this asset-light approach would allow it to develop state-of-the-art driving tech. By selling software to car and parts makers, it not only brings in income but also sources mountains of data, including how and when humans intervene, to train its codes at relatively low costs.
The company declined to share who its clients are but said they include top carmakers and Tier 1 suppliers in China and overseas. There won’t be many of them because a “partnership” in the auto sector demands deep, resource-intensive collaboration, so less is believed to be more. What we do know is Momenta counts Daimler AG as a backer. It’s also the first Chinese startup that the Mercedes-Benz parent had ever invested in, though Cao would not disclose whether Daimler is a client.
“Say you operate 10,000 autonomous cars to reap data. That could easily cost you $ 1 billion a year. 100,000 cars would cost $ 10 billion, which is a terrifying number for any tech giant,” Cao said. “If you want to acquire seas of data that have a meaningful reach, you have to build a product for the mass market.”
Highway Pilot, the semi-autonomous solution that was controlling our car, is Momenta’s first mass-produced software. More will launch in the coming seasons, including a fully autonomous parking solution and a self-driving robotaxi package for urban use.
In the long run, the startup said it aims to tackle inefficiencies in China’s $ 44 billion logistics market. People hear about warehousing robots built by Alibaba and JD.com, but overall, China is still on the lower end of logistics efficiency. In 2018, logistics costs accounted for nearly 15 percent of national gross domestic product. In the same year, the World Bank ranked China 26th in its logistics performance index, a global benchmark for efficiency in the industry.
Cao Xudong, co-founder and CEO of Momenta / Photo: Momenta
Cao, an unassuming CEO, raised his voice as explained the company’s two-legged strategy. The twin approach forms a “closed loop,” a term that Cao repeatedly summoned to talk about the company’s competitive edge. Instead of picking between the presence and future, as Waymo does with Level 4 — a designation given to cars that can operate under basic situations without human intervention — and Tesla with half-autonomous driving, Momenta works on both. It uses revenue-generating businesses like Highway Pilot to fund research in robotaxis, and the sensor data collected from real-life scenarios to feed models in the lab. Results from the lab, in turn, could soup up what gets deployed on public roads.
Human or machine
During the 40-minute ride in midday traffic, our car was able to change lanes, merge into traffic, create distance from reckless drivers by itself except for one brief moment. Toward the end of the trip, our driver decided to grab the wheel for a lane change as we approached a car dangerously parked in the middle of the exit ramp. Momenta names this an “interactive lane change,” which it claims is designed to be part of its automated system and by its strict definition is not a human “intervention”.
“Human-car interaction will continue to dominate for a long time, perhaps for another 20 years,” Cao noted, adding the setup brings safety to the next level because the car knows exactly what the driver is doing through its inner-cabin cameras.
“For example, if the driver is looking down at their cellphone, the [Momenta] system will alert them to pay attention,” he said.
I wasn’t allowed to film during the ride, so here’s some footage from Momenta to give a sneak peek of its highway solution.
Human beings are already further along the autonomous spectrum than many of us think. Cao, like a lot of other AI scientists, believes robots will eventually take over the wheel. Alphabet-owned Waymo has been running robotaxis in Arizona for several months now, and smaller startups like Drive.ai are also offering a similar service in Texas.
Despite all the hype and boom in the industry, there remains thorny questions around passenger safety, regulatory schema and a host of other issues for the fast-moving tech. Uber’s fatal self-driving crash last year delayed the company’s future projects and prompted a public backlash. As a Shanghai-based venture capitalist recently suggested to me: “I don’t think humanity is ready for self-driving.”
The biggest problem of the industry, he argued, is not tech-related but social. “Self-driving poses challenges to society’s legal system, culture, ethics and justice.”
Cao is well aware of the contention. He acknowledged that as a company with the power to steer future cars, Momenta has to “bear a lot of responsibility for safety.” As such, he required all executives in the company to ride a certain number of autonomous miles so if there’s any loophole in the system, the managers will likely stumble across it before the customers do.
“With this policy in place, the management will pay serious attention to system safety,” Cao asserted.
Momenta’s new headquarters in Suzhou, China / Photo: Momenta
In terms of actually designing the software to be reliable and to trace accountability, Momenta appoints an “architect of system research and development,” who essentially is in charge of analyzing the black box of autonomous driving algorithms. A deep learning model has to be “explainable,” said Cao, which is key to finding out what went wrong: Is it the sensor, the computer, or the navigation app that’s not working?
Going forward, Cao said the company is in no rush to make a profit as it is still spending heavily on R&D, but he assured that margins of the software it sells “are high.” The startup is also blessed with sizable fundings, which Cao’s resume certainly helped attract, and so did his other co-founders Ren Shaoqing and Xia Yan, who were also alumni of Microsoft Research Asia.
As of last October, Momenta had raised at least $ 200 million from big-name investors including GGV Capital, Sequoia Capital, Hillhouse Capital, Kai-Fu Lee’s Sinovation Ventures, Lei Jun’s Shunwei Capital, electric vehicle maker NIO’s investment arm, WeChat operator Tencent and the government of Suzhou, which will house Momenta’s new 4,000 sq-meter headquarters right next to the city’s high-speed trail station.
When a bullet train speeds past Suzhou, passengers are able to see from their windows Momenta’s recognizable M-shape building, which, in the years to come, might become a new landmark of the historic city in eastern China.