The Scenario Builder support package provides tools for you to generate high-fidelity simulation scenarios from recorded real-world driving data. Using advanced AI models, the package extracts road networks, actor trajectories, and critical driving events from GPS/IMU, camera, lidar, and object list data. You can customize scenario generation workflows by tuning event detection parameters, selecting relevant sensors or sensor combinations, and synchronizing multi-sensor data streams. You can also use generated scenarios in RoadRunner and Driving Scenario Designer or export to ASAM-standard formats such as OpenSCENARIO®, OpenDRIVE®, and OpenCRG®.
Scenario Builder lets you:
- Reconstruct scenes with lanes, trees, and buildings
- Create scenarios from vehicle trajectories
- Detect and extract critical driving events
- Generate custom driving assets from images for simulation
- Import and analyze driving data
Reconstruct Scenes with Lanes, Trees, and Buildings
Build simulation-ready road environments directly from recorded driving data. Extract complete road networks—including lanes, junctions, curvature, and banking—from camera, lidar, and lane detection sources. Reconstruct static elements like trees, buildings, and traffic signs using real-world sensor data or public aerial lidar like USGS. Refine maps based on real-world elevation profiles, lane counts, and lane widths—significantly reducing manual creation.
Create Scenarios from Vehicle Trajectories
Programmatically generate scenarios by extracting ego and target actor trajectories from GPS, camera, lidar, and object list data. Correct the ego vehicle’s pose and trajectory using IMU and lane position information to ensure accurate scenario reconstruction—no manual trajectory creation required.
Detect and Extract Critical Driving Events
Detect important driving behaviors like cut-ins, lane changes, turns, braking, and close crossings directly from recorded sensor data. Adjust event detection parameters to extract variations, such as aggressive cut-ins, and define custom event triggers, enabling you to build edge-case scenarios database programmatically.
Generate Custom Driving Assets from Images for Simulation
Programmatically create custom vehicles and static objects (trees, buildings, pylons) from a single high-resolution image. Transform, scale, and position assets into RoadRunner Scenario coordinates to simulate real-world environments with custom objects—ready for scenario testing and development.
Import and Analyze Driving Data
Import real-world driving data from GPS, camera, lidar, trajectory, and object list recordings stored in Rosbag, tabular, or MAT formats. Align and synchronize multi-sensor data streams using visual cues and calibration information, such as lane positions or satellite imagery. Overlay detected lanes and actor tracks on raw sensor data for visual validation, ensuring accuracy before scenario generation.
Download Support Package
Ready to Talk?
Let us know how we can help you.
