Uber ATG R&D
We are leading the industry with cutting-edge research for self-driving cars in broad areas of computer vision, machine learning, and robotics, publishing our findings in top conferences and journals.
Perception and prediction
ATG’s work perceives and understands the dynamic environment around a self-driving car through a multi-sensor setup and uses this sensor information to predict possible future states of the world.
Given the observations and the predictions of the dynamic actors around the self-driving vehicle, as well as the static constructs provided by HD maps, the problem of motion planning is to efficiently solve for the trajectory the self-driving vehicle will follow.
Simulation facilitates realistic scenario evaluation and training for the entire self-driving stack, allowing for scalable research and safety analysis of self-driving algorithms.
Efficient and effective multi-sensor approaches help localize the self-driving vehicle within the HD map in real time. Our research focuses on LiDAR point clouds, camera imagery and other sources of information to boost the robustness of localization algorithms.
Core AI tech
The fundamentals of machine learning science such as sparse and efficient operations for real-time inference, theoretical analysis of the robustness and safety of neural networks, stochastic and non-linear optimization, learning in the presence of noisy biased data, and more.
Publications and Articles
We publish novel state-of-the-art algorithms for self-driving through top conferences and journals in the areas of computer vision, machine learning, and robotics.
Making self-driving cars a reality is being part of history in the making. We’re looking for world-class researchers, research engineers, and software developers who prioritize safety and want to join our mission to bring safe, reliable self-driving transportation to everyone, everywhere.