Uber revealed on Wednesday a prototype car that it plans to use to scoop up real-world driving data for its growing roster of autonomous vehicle partners, including Avride, Waymo, and WeRide.
The vehicle is not some radical design. Rather, it’s a Hyundai Ioniq 5 fitted with an incredible number of sensors on the top and sides, as the company first told TechCrunch back in January. The sensor-laden vehicle may not look particularly groundbreaking, but it does mark a few milestones for the company.
This is the first vehicle Uber has assembled itself (with help from a partner) since the company sold its autonomous vehicle division to Aurora in 2020. It also represents progress on Uber’s new AV Labs division, which launched earlier this year to use sensor-equipped Uber cars to collect and then share data with its 30-and-counting autonomous vehicle technology partners.
Uber said Wednesday that it plans to roll out 500 of these kitted-out Hyundai EVs globally this year. That fleet will be able to collect “2 million miles per month of high-fidelity data” for robotaxis. Uber expects 50 of these vehicles to be on the road by the summer.
The Ioniqs are fitted with 14 cameras, eight solid-state lidar sensors, and nine radars through a partnership with Roush Performance, which will handle the vehicle retrofits. All of that data will be routed through Nvidia’s Dual Drive Thor autonomous vehicle computer. Uber appears open to changing that sensor suite, noting that it will continue to update it as its partners’ needs evolve.
This isn’t just a transfer of raw data, however. Uber said its aim is to develop the world’s most geographically diverse set of training data specifically geared towards autonomous driving. If successful, this dataset will give its AV partners a 360-degree, time synchronized stitched view that can be used to train self-driving software.
The company already has a head start. Uber told TechCrunch it has collected data from thousands of vehicles equipp …