Zoox brings its self-driving Toyotas to Seattle


Seattle residents can start keeping their eyes peeled for matte black Toyota Highlanders, coated with a wide array of sensors on their rooftops, plying the streets of their drizzly city. These retrofit SUVs belong to Zoox, an Amazon-owned self-driving car company that one day wants to run autonomous taxi services, like Uber, but without a driver behind the wheel.

Self-driving cars are nothing new, of course, even if Zoox’s vision is particularly brilliant. The company is working on an autonomous vehicle which, in general, belongs to a category of vehicles that tend to be jokingly compared to rotary toasters. In the case of Zoox, their future taxi will have four seats and will be able to drive two-way, since it has neither front nor rear like a normal car does. It also boasts four-wheel steering and paints any place for a driver. It has no official name just yet. The company likes to compare it to a carriage, because the four passengers will sit facing each other.

In Seattle, the company plans to open an office next year and is also bringing a double-digit fleet of Toyota Highlanders to the streets (not futuristic cabs). In fact, the company has already discreetly made some exploratory moves on the streets of the city that is home to their parent company, Amazon.

“We actually first sent a Highlander to Seattle in late 2019,” says Jesse Levinson, co-founder and CTO of the company. With these SUVs, the company can use sensors on them – devices such as lidar units and cameras – to create highly detailed maps of the urban environment, a key prerequisite for a self-driving car. “We were able to build a map of downtown Seattle very quickly, and then on our next trip we were able to drive fully autonomously in downtown Seattle,” says Levinson. “Day, night and rain, and it worked really, really well.”

“We did it very quietly,” he adds.

A Zoox L3 vehicle. The company never plans to use them for running a commercial taxi service. Zoox

Looking for wet skies

Now, society doesn’t shut up about it, announcing today they are officially expanding to Seattle. They have already tested their Highlander vehicles in other locations: two points in California and Las Vegas.

Levinson notes that Seattle will bring some challenges for these Toyota Highlanders, which they brag about level three autonomy and have safe drivers behind the wheel. One of these challenges will be the famous drizzle of the city. “We’re looking for some rain to test our sensors,” Levinson says. “We really designed our vehicle to handle inclement weather, but we don’t get a ton in San Francisco and even less in Las Vegas, so having frequent rains will be really helpful for that purpose. ”

He also notes that simply being in a new position will help shape the company’s software. “Getting more data diversity is good for machine learning algorithms,” he says.

[Related: What will it take for humans to trust self-driving cars?]

At a point yet to be announced, Zoox hopes to open a robo-taxi service for customers who would accept one of their new carriage-like vehicles, something like one of their rivals, Waymo, is operational at the moment in the Phoenix area with driverless Chrysler Pacificas. The new generation Zoox taxis sport lidar sensors at their corners, as well as other sensors, such as cameras, radar units, and an IMU or inertial measurement unit. Additionally, the vehicle’s roof will feature a sunroof and hundreds of LED lights to create an atmosphere for future passengers. (Another company, Cruise, is working on a similar vehicle, also in the wheeled toaster category, called Origin.)

Meanwhile, the Toyota Highlanders are the more pragmatic vehicles on the company’s stepping stone, which they will use to conduct mapping and autonomous driving in Seattle. The company does not intend to use the Toyotas as part of the future ride-hailing service, but Levinson says the sensor setup on the Highlanders is similar enough to the next-generation vehicle that the data collected with the SUVs is still useful. “They have the same sensors and are in nearly identical geometric configurations,” he says.

Why is this Amazon-owned company bringing its autonomous vehicles to Seattle?
The new generation vehicle that Zoox plans to use for a future robo-taxi service. It has no space for a driver and its four occupants sit facing each other. Zoox

Modern cartography

For a self-driving car to know where it is in the urban landscape, it needs a detailed three-dimensional map to refer to and on-board sensors to sense what’s around. The map, of course, is very different from what a human driver navigates when launching a navigation app.

“The maps we use in self-driving cars are completely different and are much more detailed, much more faithful,” says Taylor Arnicar, Zoox’s technical personnel program manager. “And even the things that humans might consider important have nothing to do with the self-driving car itself: the self-driving car doesn’t care what the street name is, actually.” (Instead, the car uses numbers to figure out which street it is on.)

But before the vehicles can use those maps, the company has to build them, which they do using one of those Highlanders. “We will be driving those vehicles into a new area that we want to explore in the future,” says Taylor. This allows them to connect the information they need to create maps.

[Related: Kitty Hawk’s electric airplane will fly you around—with no pilot]

Interestingly, the process works best if the terrain is varied, instead of including bridge or tunnel spans that contain very repetitive features and therefore could be confusing. “These mapping systems are most powerful when they operate in areas with many unique 3D structures in the world,” he says. Human eyes may have trouble determining their position on a desolate stretch of road in a featureless desert, or on a long bridge span that looks just like the front, and the same goes for a robotic car.

A tunnel in Seattle, as captured by lidar. Zoox

The “phase of reality”

Of course, Zoox isn’t the only player in the autonomous vehicle space. Other noteworthy competitors include Aurora, Argo AI, Cruise, motivational, and of course Waymo, which has already built a driverless taxi service in the Phoenix, Arizona area. Waymo is also directing some rides in the San Francisco area, on a smaller scale and using electric jaguars.

Computer-driven cars are certainly still flawed and tend to run into obstacles that may not mislead a human driver. (Then again, they never get tired or distracted.) In San Francisco, some Waymo vehicles have appeared mysteriously in a dead end street. In Phoenix, Waymo has he reportedly struggled with issues such as left turns (Which ones are they difficult even for self-driving cars and normal cars) and even puddles.

[Related: How Waymo is teaching self-driving cars to deal with the chaos of parking lots]

The industry’s goal isn’t just to build robot-taxis – other industries involve autonomous trucking, self-driving shuttles, and even the kind of driver assistance features designed for regular cars, like GM’s Super Cruise.

Currently, the self-driving auto industry has reached something of a “reality phase,” says Raj Rajkumar, who heads the Metro21: Smart City Institute at Carnegie Mellon University. He noted a “huge hype cycle” that peaked around 2018 and then an industry “lull” in 2019.

“I think he’s starting to recover,” he says. “Right now, the hype is much more subdued.”


Please enter your comment!
Please enter your name here