Pick on Simulation (test)
5.11Trajectory RMSE (deg)Ours (PP)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Ours (PP)Pairing Protocol=Paired (Matching Algorithm), Trainable Weights=940K2026.03 | 5.11 | 7.04 | 8.2 | |
| DP-ModeTrainable Weights=67M, Backbone=Diffusion Policy2026.03 | 8.75 | 13.44 | 4.2 | |
| DP-2HeadTrainable Weights=67M, Backbone=Diffusion Policy2026.03 | 16.46 | 20.56 | — | |
| DP-DualTrainable Weights=135M, Backbone=Diffusion Policy2026.03 | 25.36 | 20.52 | — | |
| Ours (UP)Pairing Protocol=Unpaired (Randomly Paired), Trainable Weights=940K2026.03 | 26.21 | 14.12 | 3.4 |