| --- |
| license: apache-2.0 |
| --- |
| |
| # ReactSim-Bench: Benchmarking Reactive Behavior World Model Simulation in Autonomous Driving |
|
|
| ReactSim-Bench is the first benchmark for systematicly evaluating the reactive capability of behavior world models in autonomous driving. It contains: |
|
|
| - Reactive closed-loop protocol with decoupled control. In ReactSim-Bench, The behavior world model controls the surrounding agents, while the autonomous vehicle (AV) are controled by its own policy instead of the world model. |
| - Customed AV behaviors beyond the log. ReactSim-Bench contains 2,636 scenarios with AV behaviors that differ from the log and create reactive pressure on surrounding agents. They are grouped into three categories: longitudinal,directional, and lateral deviations. |
| - Safety and feasibility metrics. ReactSim-Bench evaluates Agent-AV safety, agent-agent safety, map compliance, driving-direction compliance, and kinematic feasibility. |
| - Multiple baselines. We implement the Transformer-based (MTR), diffusion-based (CTG,VBD), and next-token-prediction-based (SMART, catk, Trajtok) behavior world models on ReactSim-Bench as baselines. |
|
|
| This dataset is based on nuPlan. Each file is named as `<map_name>_<scene_token>_<sample_token>.pkl`. This dataset only contains the customed behaviors of autonomous vehicles, without any information in original nuPlan scenarios (states of agents, HD maps or sensors). |
| For the nuPlan scenarios, please download from its official website. |
|
|
| For more information, please visit our GitHub repository: https://github.com/Thinklab-SJTU/ReactSim-Bench. |
