Remote core developer on the CARLA Autonomous Driving Leaderboard team.
CARLA Leaderboard 2.0
Remote core development work for the CARLA Autonomous Driving Leaderboard, focused on new scenarios, map testing, route validation, and performance optimization.
CARLAsimulation testingbenchmarkingscenario validation
- Problem
- Autonomous driving benchmarks need standardized routes, maps, and testing scenarios so systems can be evaluated and compared under repeatable simulation conditions.
- Method
- Contributed to CARLA Leaderboard 2.0 by creating and validating new testing scenarios, introducing new-map tests, optimizing system performance, and designing and validating routes for Town13.
- Results
- Supported four global challenges from 2021 to 2024, contributed 15 merged pull requests to the official carla-simulator GitHub repository, designed and validated all 10 Town13 challenge routes, and developed 9 new testing scenarios.
This project is the strongest open-source benchmark contribution in the source material. It connects Songyang Yan’s research focus with practical simulator infrastructure used by CARLA’s global autonomous driving challenges.