This vehicle team will play the role of collecting real traffic data, serving the training and completion of automatic driving systems of many technology partners.
The model chosen is the Hyundai Ioniq 5, which has been improved with a series of sensors mounted on the roof and on both sides of the body to record the traffic environment from many different angles.
Although not a breakthrough design, this is still a notable milestone for Uber. This is the first car that the company has self-developed and installed a sensor system since selling its self-driving car division to Aurora (a famous American startup specializing in self-driving car technology) in 2020.
At the same time, the project also shows the first steps of AV Labs, a research unit established by Uber earlier this year to support partners in developing self-driving car technology.
According to the plan, Uber will deploy 500 vehicles globally in 2026. The company said the fleet can generate about 2 million miles of high-quality data per month. In the early stages, about 50 vehicles will start rolling in the summer of this year.
The vehicles are equipped with 14 cameras, 8 solid lidar sensors and 9 radars. The upgrade is carried out with partner Roush Performance (a high-performance car tuning and component manufacturing company).
All data from the sensors will be processed through Nvidia's Thor Dual Drive computer platform, allowing real-time data recording and synchronization.
Uber said its goal is not only to collect data but also to build one of the most diverse self-driving car training datasets in the world.
Thanks to the combination of images, lidar and radar, partners can access data sources simulating the traffic environment with a 360-degree viewing angle, helping to improve the recognition and decision-making capabilities of AI systems.
This data warehouse will be shared with more than 30 Uber's self-driving car technology partners, including Waymo, WeRide and Avride. The company expects a large-scale real-world data source to help shorten development time and improve the safety of self-driving vehicles.
In fact, Uber has built a significant data platform in recent years. The company said it has collected information from thousands of camera-mounted vehicles operating in many cities around the world through operating partners.
In addition, data from hundreds of Lucid Air electric vehicles in the US and Europe have also been recorded in the past two years.
AV Labs is currently compiling and analyzing these data sources, and is preparing to add a large amount of data from the new Hyundai Ioniq 5 car fleet.
The above move is part of Uber's long-term strategy to expand its presence in the field of automated transportation.
In February, the company established the Uber Autonomous Solutions division to manage activities related to self-driving taxis, automatic trucks and delivery robots.
Strong investment in data shows that Uber is betting on the role of a connecting platform between self-driving technology developers and transportation networks in the future.
