SimBEV2X: A Large-Scale Dataset and Data Generation Tool for Multi-Task Vehicle-to-Everything Cooperative Perception
2026-07-27 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMachine LearningMultiagent SystemsRobotics
AI summaryⓘ
The authors created SimBEV2X, a tool that uses a driving simulator to make a very large and detailed dataset for studying how self-driving cars share information to see better together. This dataset includes lots of sensor data, maps, and labeled objects from many vehicles and roadside units driving in different scenarios. Because collecting such real-world data is very expensive, their synthetic dataset helps researchers develop and test new vehicle communication methods. They also introduced a new model called CoBEVFusion that improves how vehicles combine information from others for better perception. All their tools and data are openly available online.
cooperative perceptionvehicle-to-everything (V2X)bird's-eye view (BEV) representationsynthetic datasetCARLA simulatormulti-modal sensor data3D bounding boxesmulti-agent feature aggregationaxial attentionautonomous vehicles
Authors
Goodarz Mehr, Sepideh Gohari, Montasir Abbas, Azim Eskandarian
Abstract
Cooperative perception through vehicle-to-everything (V2X) communication can overcome the inherent physical limitations of individual autonomous vehicles, such as occlusions and limited sensor range. However, the development of robust V2X algorithms, particularly those relying on unified spatial representations like bird's-eye view (BEV) representation, is hampered by the lack of large-scale, multi-modal, multi-task datasets. Moreover, collecting and annotating a large set of synchronized, real-world multi-agent data is prohibitively expensive. This has resulted in a landscape where existing V2X datasets are notably limited in both size and scope. To overcome this, we introduce SimBEV2X, an advanced synthetic data generation tool built on the CARLA simulator. SimBEV2X automatically creates randomized driving scenarios to collect multi-modal sensor data alongside various types of ground truth including 3D bounding boxes with unique track IDs, HD map information, BEV segmentation maps, and semantic occupancy voxel grids from both vehicles and RSUs. We also present the SimBEV2X dataset, the largest V2X perception dataset to date. The dataset comprises 258 scenes, each involving up to 8 connected vehicles and up to 4 RSUs across a variety of road networks. The SimBEV2X dataset is an order of magnitude larger than existing V2X datasets and contains 102,200 frames, 588,520 lidar point clouds, more than 3 million images, over 27 million bounding boxes, and a comprehensive set of other annotations. Finally, we establish a strong baseline on the SimBEV2X dataset using CoopDet3D and propose CoBEVFusion, a novel architecture that combines CoopDet3D with fused axial attention (FAX) for context-aware multi-agent feature aggregation, resulting in superior performance. SimBEV2X, the SimBEV2X dataset, and CoBEVFusion are available at https://simbev2x.org and https://github.com/GoodarzMehr/SimBEV2X.