Autonomous Driving Datasets
Overview
Autonomous driving datasets capture road scenes, road users, vehicle motion, and driving decisions so systems can learn and be tested in varied conditions. They include naturalistic recordings from cameras, LiDAR, radar, vehicles, and drones; simulated or generated scenes; and annotations ranging from object labels and trajectories to maps, risks, and language-based reasoning. Benchmarks use these data to compare perception, forecasting, localization, and planning, including in long-tail scenarios, adverse weather, and cooperative settings. A central challenge is building data that is diverse, well-labeled, and realistic enough to reveal failures beyond familiar routes and short, open-loop tests, while keeping collection and annotation practical.
Latest papers
141 papers11 papers in the last four weeks, down 39% on the four weeks before. 0.1% of all new papers.