Robot Navigation on Unitree Go2 Narrow Corridor & Cluttered Office (Real-world deployment)
70Success RateReinforcement Learning Fine-Tuning for Diffusion Navigation Policies
Evaluation results
| Method | Links | |
|---|---|---|
| Reinforcement Learning Fine-Tuning for Diffusion Navigation PoliciesState=RL-FT2026.03 | 70 | |
| NavDPState=Pretrained2026.03 | 50 |