KLABCMU.GITHUB.IO

The All-in-One Drive (AIODrive) is a large-scale comprehensive perception dataset for autonomous driving developed by Carnegie Mellon University in order to innovate multi-sensor multi-task perception systems for autonomous driving. Though various driving datasets have been released to advance perception systems, these datasets often have different objectives and provide different annotations, sensors and environmental variations such as 3D segmentation (SemanticKITTI), rich maps (Argoverse), radar sensing (nuScenes), high-speed driving (A*3D), adverse weather and large-scale data (Waymo). In contrast, our goal is to build a comprehensive dataset that combines the strengths of all existing perception datasets, in order to further innovate multi-sensor multi-task perception systems. To that end, our AIODrive dataset provides the most comprehensive sensing modalities, support annotations for all mainstream perception tasks and collect diverse driving environmental variations...