Six years after scrapping customized cars for the self-driving market, Uber is back—if not in the way you’d expect. The ridesharing giant has revealed a prototype version of Hyundai’s Ioniq 5 that will be used to gather self-driving data for partners like Waymo and WeRide.
The customized EV adds eight lidar (laser-based) sensors, nine radar sensors, and 14 cameras through an alliance with the tuning company Roush Performance. One of NVIDIA’s Dual Drive Thor computers will process the collected data.
A full 500 examples will start driving worldwide this year, Uber says, with the first 50 hitting streets this summer.
The company hopes to gather about two million miles of “high-fidelity” information every month, and to produce the most diverse training dataset possible for autonomous vehicles. Partners can use the content to get a fuller understanding of how a self-driving car navigates or reacts to unexpected situations.
Why Uber is launching its own cars again
It wants to be your source for self-driving data
Uber VW ID. Buzz robotaxis driving in Los Angeles.Credit: Uber/MOIA
Uber hasn’t operated custom vehicles since 2020, when it sold its autonomous driving unit to Aurora Innovation. The company initially hoped to operate a complete service with its own cars, but those plans ground to a halt after a 2018 incident where a test car struck and killed a pedestrian.
The company instead switched to offering ride hailing services for partners that were potential competitors, such as Waymo, WeRide, and Nuro. You might not sit in an Uber car, but you’re using its app to book trips.
Uber expanded its ambitions earlier this year when it launched an AV Labs division meant to collect and share data. Ideally, the company becomes even more indispensable as robotaxi operators depend on it to scoop up real-world driving knowledge they can’t get through their own cars or simulations.
Related
Waymo starts offering rides in its new robotaxi van—a ‘living room on wheels’
The Ojai also debuts a new self-driving AI.
Leading from behind the scenes
With this approach, Uber theoretically succeeds even if it never deploys its own robotaxis. The larger a driverless brand grows, the more likely it is to need data to improve its efficiency and safety. While larger companies like Waymo and Volkswagen are often large enough to rely on their own know-how, this helps them accelerate their rollouts—you might be hailing a driverless ride that much sooner as a result.
Source: Uber
