RACECAR LiDAR Datasets

RACECAR multislow_poli

This dataset contains a subset of 4D point cloud data from LiDAR sensors collected from fully autonomous and self-driving Indy race cars that raced in the Indy autonomous challenge. The dataset is in nuScenes format and is divided into 7,150 sweeps and 1,199 samples, which contain fused sensor data from 3 LiDARs equipped on the vehicle. This dataset’s scenario is the PoliMove team’s Multi-Agent Slow on the LVMS racetrack. Each .pcd file contains 4-dimensional data: (x,y,z) coordinates in 3D space and an intensity value for each point. The dataset is available at huggingface and kaggle. This dataset was used for the research work Parallel Neural Computing for Scene Understanding from LiDAR Perception in Autonomous Racing.

RACECAR LiDAR to BEV Map image labeled

This dataset contains processed and converted LiDAR point clouds(3D) into RGB bird’s-eye view images(2D) of the Racecar Dataset’s multi-slow-poli race scenario. With labels for object (other race cars on the track) detection and trajectory generation/planning. The images are segmented scenes along space-time dimensions, with each segmentation covering a scene history of 15 frames. The dataset is available at huggingface and kaggle.