Uber has announced the establishment of a new division called Uber AV Labs, aimed at gathering driving data to support its autonomous vehicle partners, as reported by TechCrunch. This move signifies Uber’s shift towards offering data services rather than developing its own self-driving vehicles, following a tragic incident in 2018.
The company, with over 20 autonomous vehicle partners including Waymo and Lucid Motors, recognizes the critical need for real-world driving data in refining self-driving technologies. By equipping its vehicles with sensors for data collection, Uber aims to assist partners in addressing complex edge cases that traditional rule-based systems struggle to navigate.
Autonomous vehicle development is evolving towards reinforcement learning, emphasizing the value of extensive driving data for training AI systems. Companies like Waymo have faced challenges, such as vehicles illegally passing stopped school buses, highlighting the necessity for diverse and abundant data sets.
Uber’s Chief Technology Officer, Praveen Neppalli Naga, emphasized the significance of volume in tackling extreme driving scenarios, underscoring the limitations posed by a company’s fleet size on data collection. The initiative reflects a collaborative approach within the autonomous driving industry to leverage shared data resources for enhanced system training and performance.
Source: TechCrunch