One-step Prediction on DDR (Differential-Drive Robot) under unmodeled dynamics
1.1Mean ErrorRBF (DZ-CS)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RBF (DZ-CS)Dictionary Type=RBF, Adaptation Strategy=dead-zone with covariance skipping2026.06 | 1.1 | 3.5 | 3 | 97.74 | |
| Offline RBFDictionary Type=RBF, Adaptation Strategy=Offline2026.06 | 1.3 | — | 7.6 | — | |
| Poly (DZ-CS)Dictionary Type=Polynomial, Adaptation Strategy=dead-zone with covariance skipping2026.06 | 1.9 | 1.9 | 5.3 | 3.24 | |
| PolyDictionary Type=Polynomial, Adaptation Strategy=Recursive2026.06 | 2 | 1.9 | 5.3 | — | |
| Poly (DZ-KS)Dictionary Type=Polynomial, Adaptation Strategy=dead-zone with full Koopman skipping2026.06 | 2 | 1.9 | 5.3 | 3.02 | |
| Poly (CT)Dictionary Type=Polynomial, Adaptation Strategy=constant-trace normalization2026.06 | 4.5 | 9 | 6.4 | — | |
| RBF (DZ-KS)Dictionary Type=RBF, Adaptation Strategy=dead-zone with full Koopman skipping2026.06 | 4.6 | 2.4 | 3 | 95.8 | |
| RBFDictionary Type=RBF, Adaptation Strategy=Recursive2026.06 | 5.3 | 2.5 | 2 | — | |
| RBF (CT)Dictionary Type=RBF, Adaptation Strategy=constant-trace normalization2026.06 | 5.6 | 6.6 | 3 | — |