Robot Trajectory Optimization on 16 Robot Experiments Tabletop (test)
13T1BVLA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VLAModel Type=Vision-Language-Action (GR00T), Training=Fine-tuned2025.12 | 13 | 31 | 42.03 | 42.28 | |
| DPModel Type=Diffusion Policy2025.12 | 14 | 29 | 47.81 | 48.57 | |
| HumanMode=Teleoperated demonstrations2025.12 | 62 | 191 | — | — | |
| Risk (Semantic-Metric Bayesian Risk Fields)Optimizer=Classical trajectory optimizer, Viability Threshold=alpha = 0.12025.12 | 134 | 196 | 41.1 | 44.23 |