ad-experiment-design

You can design autonomous driving safety experiments by running the `design_experiment` command with parameters for method, environment, and speed. This framework manages simulation configurations, dataset selection, and statistical analysis.

You can use real-world datasets like nuScenes, Waymo, KITTI, and FARS for robustness testing and safety evaluations. These datasets help validate autonomous driving models under diverse real-world conditions.

Yes, the framework integrates with CommonRoad, CARLA, and SUMO for comprehensive simulation studies. It provides configuration tools to set up these AD simulators for autonomous driving safety research.

The framework provides structured statistical analysis methods for autonomous driving safety experiments, including variable management and test matrix creation. This ensures robust evaluation of simulation and real-world testing results.

You set up a test matrix by defining experiment variables like method, environment, and speed using the `design_experiment` command. This creates a structured approach for evaluating autonomous driving safety across different simulation and real-world scenarios.

Yes, you can utilize the FARS dataset alongside nuScenes, Waymo, and KITTI for robustness testing and safety evaluations. These real-world datasets help validate autonomous driving models against actual accident statistics.