Simulation using realistic driving scenarios and sensor models is a crucial part of testing automated driving algorithms. Automated Driving Toolbox™ enables you to create driving scenarios with synthetic sensor data. The toolbox provides these simulation environments to test automated driving algorithms. These environments have their uses, and one environment is not a replacement for the other.
In the cuboid simulation environment, vehicles and other actors in the scenario are represented as simple box shapes, or for lidar applications, as polygon meshes. Use this environment to rapidly author scenarios, generate detections using low-fidelity radar and camera sensors, and test controllers and tracking and sensor fusion algorithms in both MATLAB® and Simulink®. To get started authoring a scenario, use the Driving Scenario Designer app.
You can also import road networks from the ASAM OpenDRIVE® file format, edit the imported road networks, and build scenarios. Then, you can export driving scenarios to the ASAM OpenDRIVE and ASAM OpenSCENARIO® formats.
In the Unreal Engine® simulation environment, scenarios are rendered using the Unreal Engine from Epic Games®. Use this environment to visualize scenarios using more realistic graphics; to generate high-fidelity radar, camera, and lidar sensor data; and to test perception-in-the-loop systems. This environment is available in Simulink and runs on Windows® only. To learn more, see Unreal Engine Simulation for Automated Driving.
In the RoadRunner scenario simulation environment, scenes and scenarios are created using RoadRunner and RoadRunner Scenario. Automated Driving Toolbox provides tools to programmatically manage scenes and scenarios and simulate scenarios in RoadRunner with actors modeled in MATLAB and Simulink. To learn more, see Overview of Simulating RoadRunner Scenarios with MATLAB and Simulink.
Automated Driving Toolbox also provides these support packages that enable you to build scenarios from recorded sensor data and generate multiple variants of a seed scenario to perform large-scale testing.
Scenario Builder for Automated Driving Toolbox support package provides tools to automatically build and visualize virtual scenarios from real-world data recorded using various sensors such as GPS, IMU, camera, and lidar. To learn more, see Overview of Scenario Generation from Recorded Sensor Data.
Scenario Variant Generator for Automated Driving Toolbox support package provides tools to modify scenario parameters and create multiple scenario variants from a seed scenario. Use these scenario variants to test automated driving applications as per European New Car Assessment Programme (Euro NCAP®) test protocols. To learn more, see Overview of Scenario Variant Generation.
- Create Driving Scenario Interactively and Generate Synthetic Sensor Data
- Customize Unreal Engine Scenes for Automated Driving
- Import ASAM OpenDRIVE Roads into Driving Scenario
- Export Driving Scenario to ASAM OpenSCENARIO File
- Overview of Scenario Generation from Recorded Sensor Data
- Overview of Scenario Variant Generation
- Cuboid Scenario Simulation
Test automated driving algorithms using cuboid-based scenarios and detections from radar, camera, and lidar sensor models
- Unreal Engine Scenario Simulation
Develop, test, and visualize the performance of driving algorithms in a simulated environment using the Unreal Engine from Epic Games
- RoadRunner Scenario Simulation
Design scenarios for simulating and testing automated driving systems with RoadRunner Scenario
- Scenario Generation and Variation
Generate scenario from recorded real-world sensor data and create scenario variations from seed scenario