Archives

self-driving cars

The rise of robotaxis in China

AutoX, Momenta and WeRide took the stage at TC Sessions: Mobility 2021 to discuss the state of robotaxi startups in China and their relationships with local governments in the country.

They also talked about overseas expansion — a common trajectory for China’s top autonomous vehicle startups — and shed light on the challenges and opportunities for foreign AV companies eyeing the massive Chinese market.


Enterprising governments

Worldwide, regulations play a great role in the development of autonomous vehicles. In China, policymaking for autonomous driving is driven from the bottom up rather than a top-down effort by the central government, observed executives from the three Chinese robotaxi startups.

Huan Sun, Europe general manager at Momenta, which is backed by the government of Suzhou, a city near Shanghai, said her company had a “very good experience” working with the municipal governments across multiple cities.

In China, each local government is incentivized to really act like entrepreneurs like us. They are very progressive in developing the local economy… What we feel is that autonomous driving technology can greatly improve and upgrade the [local governments’] economic structure. (Time stamp: 02:56)

Shenzhen, a special economic zone with considerable lawmaking autonomy, is just as progressive in propelling autonomous driving forward, said Jewel Li, chief operation officer at AutoX, which is based in the southern city.

ISEE brings autonomy to shipping yards with self-driving container trucks

Robotaxis may still be a few years out, but there are other industries that can be transformed by autonomous vehicles as they are today. MIT spin-off ISEE has identified one in the common shipping yard, where containers are sorted and stored — today by a dwindling supply of human drivers, but tomorrow perhaps by the company’s purpose-built robotic yard truck. With new funding and partnerships with major shippers, the company may be about to go big.

Shipping yards are the buffer zone of the logistics industry. When a container is unloaded from a ship full of them, it can’t exactly just sit there on the wharf where the crane dropped it. Maybe it’s time sensitive and has to trucked out right away; maybe it needs to go through customs and inspections and must stay in the facility for a week; maybe it’s refrigerated and needs power and air hookups.

Each of these situations will be handled by a professional driver, hooking the container up to a short-haul truck and driving it the hundred or thousand meters to its proper place, an empty slot with a power hookup, long term storage, ready access for inspection, etc. But like many jobs in logistics, this one is increasingly facing a labor shortage as fewer people sign up for it every year. The work, after all, is fairly repetitive, not particularly easy, and of course heavy equipment can be dangerous.

ISEE’s co-founders Yibiao Zhao and Debbie Yu said they identified the logistics industry as one that needs more automation, and these container yards especially. “Working with customers, it’s surprising how dated their yard operation is — it’s basically just people yelling,” said Zhao.  “There’s a big opportunity to bring this to the next level.”

Two ISEE trucks without containers on the back.

Image Credits: ISEE

The ISEE trucks are not fully custom vehicles but yard trucks of a familiar type, retrofitted with lidar, cameras, and other sensors to give them 360-degree awareness. Their job is to transport containers (unmodified, it is important to note) to and from locations in the yards, backing the 50-foot trailer into a parking spot with as little as a foot of space on either side.

“A customer adopts our solution just as if they’re hiring another driver,” Zhao said. No safe zone is required, no extra considerations need to be made at the yard. The ISEE trucks navigate the yard intelligently, driving around obstacles, slowing for passing workers, and making room for other trucks, whether autonomous or human. Unlike many industrial machines and vehicles, these bring the current state of autonomous driving to bear in order to stay safe and drive as safely as possible among mixed and unpredictable traffic.

The advantage of an automated system over a human driver is especially pronounced in this environment. One rather unusual limitation of yard truck drivers is that, because the driver’s seat is on the left side of the cabin, they can only park the trucks on the left as well since that’s the only side they can see well enough. ISEE trucks have no such limitation, of course, and can park easily in either direction, something that has apparently blown the human drivers’ minds.

Overhead view of autonomous and ordinary trucks moving around a shipping yard.

Image Credits: ISEE

Efficiency is also improved through the infallible machine mind. “There are hundreds, even thousands of containers in the yard. Humans spend a lot of time just going around the yard searching for assets, because they can’t remember what is where,” explained Zhao. But of course a computer never forgets, and so no gas is wasted circling the yard looking for either a container or a spot to put one.

Once it parks, another ISEE tech can make the necessary connections for electricity or air as well, a step that can be hazardous for human drivers in bad conditions.

The robotic platform also offers consistency. Human drivers aren’t so good when they’re trainees, taking a few years to get seasoned, noted Yu. “We’ve learned a lot about efficiency,” she said. “That’s basically what customers care about the most; the supply chain depends on throughput.”

To that end she said that moderating speed has been an interesting challenge — it’s easy for the vehicle to go faster, but it needs the awareness to be able to slow down when necessary, not just when there’s an obstacle, but when there are things like blind corners that must be navigated with care.

It is in fact a perfect training ground for developing autonomy, and that’s kind of the idea.

“Today’s robots work with very predefined rules in very constrained environments, but in the future autonomous cars will drive in open environments. We see this tech gap, how to enable robots or autonomous vehicles do deal with uncertainty,” said Zhao.

ISEE co-founders Yibiao Zhao (top), Debbie Yu (left), and Chris Baker.

ISEE Founders

“We needed a relatively unconstrained environment with complex human behaviors, and we found it’s actually a perfect marriage, the flexible autonomy we’re offering and the yard,” he continued. “It’s a private lot, there’s no regulation, all the vehicles stay in it, there are no kids or random people, no long tail like a public highway or busy street. But it’s not simple, it’s complex like most industrial environments — it’s congested, busy, there are pedestrians and trucks coming in and out.”

Although it’s an MIT spinout with a strong basis in papers and computer vision research, it’s not a theoretical business. ISEE is already working with two major shippers, Lazer Spot and Maersk, which account for hundreds of yards and some 10,000 trucks, many or most of which could potentially be automated by ISEE.

So far the company has progressed past the pilot stage and is working with Maersk to bring several vehicles into active service at a yard. The Maersk Growth Fund has also invested an undisclosed amount in ISEE, and one detects the possibility of an acquisition looming in the near future. But the plan for now is to simply expand and refine the technology and services and widen the lead between ISEE and any would-be competitors.

AI pioneer Raquel Urtasun launches self-driving technology startup with backing from Khosla, Uber and Aurora

One of the lingering mysteries from Uber’s sale of its Uber ATG self-driving unit to Aurora has been solved.

Raquel Urtasun, the AI pioneer who was the chief scientist at Uber ATG, has launched a new startup called Waabi that is taking what she describes as an “AI-first approach” to speed up the commercial deployment of autonomous vehicles, starting with long-haul trucks. Urtasun, who is the sole founder and CEO, already has a long list of high-profile backers, including separate investments from Uber and Aurora. Waabi has raised $83.5 million in a Series A round led by Khosla Ventures with additional participation from Uber, 8VC, Radical Ventures, OMERS Ventures, BDC, Aurora Innovation as well as leading AI researchers Geoffrey Hinton, Fei-Fei Li, Pieter Abbeel, Sanja Fidler and others.

Urtasun described Waabi, which currently employs 40 people and operates in Toronto and California, as the culmination of her life’s work to bring commercially viable self-driving technology to society. The name of the company —  Waabi means “she has vision” in Ojibwe and “simple” in Japanese —  hints at her approach and ambitions.

Autonomous vehicle startups that exist today use a combination of artificial intelligence algorithms and sensors to handle the tasks of driving that humans do such as detecting and understanding objects and making decisions based on that information to safely navigate a lonely road or a crowded highway. Beyond those basics are a variety of approaches, including within AI.

Most self-driving vehicle developers use a traditional form of AI. However, the traditional approach limits the power of AI, Urtasun said, adding that developers must manually tune the software stack, a complex and time-consuming task. The upshot, Urtasun says: Autonomous vehicle development has slowed and the limited commercial deployments that do exist operate in small and simple operational domains because scaling is so costly and technically challenging.

“Working in this field for so many years and, in particular, the industry for the past four years, it became more and more clear along the way that there is a need for a new approach that is different from the traditional approach that most companies are taking today,” said Urtasun, who is also a professor in the Department of Computer Science at the University of Toronto and a co-founder of the Vector Institute for AI.

Some developers do use deep neural nets, a sophisticated form of artificial intelligence algorithms that allows a computer to learn by using a series of connected networks to identify patterns in data. However, developers typically wall off the deep nets to handle a specific problem and use a machine learning and rules-based algorithms to tie into the broader system.

Deep nets have their own set of problems. A long-standing argument is that can’t be used with any reliability in autonomous vehicles in part because of the “black box” effect, in which the how and the why the AI solved a particular task is not clear. That is a problem for any self-driving startup that wants to be able verify and validate its system. It is also difficult to incorporate any prior knowledge about the task that the developer is trying to solve, like say driving. Finally, deep nets require an immense amount of data to learn.

Urtasun says she solved these lingering problems around deep nets by combining them with probabilistic inference and complex optimization, which she describes as a family of algorithms. When combined, the developer can trace back the decision process of the AI system and incorporate prior knowledge so they don’t have to teach the AI system everything from scratch. The final piece is a closed loop simulator that will allow the Waabi team to test at scale common driving scenarios and safety-critical edge cases.

Waabi will still have a physical fleet of vehicles to test on public roads. However, the simulator will allow the company to rely less on this form of testing. “We can even prepare for new geographies before we even drive there,” Urtasun said. “That’s a huge benefit in terms of the scaling curve.”

Urtasun’s vision and intent isn’t to take this approach and disrupt the ecosystem of OEMs, hardware and compute suppliers, but to be a player within it. That might explain the backing of Aurora, a startup that is developing its own self-driving stack that it hopes to first deploy in logistics such as long-haul trucking.

“This was the moment to really do something different,” Urtasun said. “The field is in need of a diverse set of approaches to solve this and it became very clear that this was the way to go.”