Continuous Robotic Locomotion Control on MuJoCo Hopper v4 (1M Steps)
30.46Normalized Final ReturnDUPO
Evaluation Results
| Method | Links | |
|---|---|---|
| DUPOMaximum observation delay (ΔTmax)=10, Training steps=1M2026.07 | 30.46 | |
| State AugmentationMaximum observation delay (ΔTmax)=10, Training steps=1M2026.07 | 23.71 | |
| DUPOMaximum observation delay (ΔTmax)=25, Training steps=1M2026.07 | 10.34 | |
| State AugmentationMaximum observation delay (ΔTmax)=25, Training steps=1M2026.07 | 7.32 | |
| DUPOMaximum observation delay (ΔTmax)=5, Training steps=1M2026.07 | 7.24 | |
| State AugmentationMaximum observation delay (ΔTmax)=5, Training steps=1M2026.07 | 6.78 | |
| State PredictionMaximum observation delay (ΔTmax)=10, Training steps=1M2026.07 | 6.41 | |
| State PredictionMaximum observation delay (ΔTmax)=25, Training steps=1M2026.07 | 4.01 | |
| State PredictionMaximum observation delay (ΔTmax)=5, Training steps=1M2026.07 | 3.19 | |
| DC/ACMaximum observation delay (ΔTmax)=5, Training steps=1M2026.07 | 1 | |
| DC/ACMaximum observation delay (ΔTmax)=10, Training steps=1M2026.07 | 1 | |
| DC/ACMaximum observation delay (ΔTmax)=25, Training steps=1M2026.07 | 1 |