Closed-loop Autonomous Driving on nuScenes reconstructed
26.76Error Rate (ER)PPO
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| PPO2026.04 | 26.76 | 10.92 | 2.65 | 0.6 | 1.08 | 0.11 | 0.39 | |
| Skill-Critic2026.04 | 30.38 | 10.68 | 2.78 | 1.87 | 1.12 | 0.08 | 0.28 | |
| RLIR2026.04 | 35.19 | 11.23 | 2.72 | 1.93 | 1.52 | 0.12 | 0.33 | |
| Q-chunking2026.04 | 39.7 | 13.9 | 1.94 | 1.65 | 0.8 | 0.09 | 0.22 | |
| RAD2026.04 | 49.24 | 12.85 | 1.68 | 2.69 | 0.68 | 0.11 | 0.19 | |
| GSDrive2026.04 | 52.97 | 13.98 | 1.56 | 3.59 | 0.52 | 0.08 | 0.11 |