Robotic motion control (Circle task) on Car-Circle
197.35RewardPPOL-vanilla
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PPOL-vanillaAttack Type=MC2026.02 | 197.35 | 25.58 | |
| TCRL-PPOLAttack Type=Random2026.02 | 179.38 | 0 | |
| TCRL-PPOLAttack Type=MC2026.02 | 178.41 | 7.58 | |
| TCRL-PPOLAttack Type=Worst-TC2026.02 | 177.19 | 2.38 | |
| PPOL-vanillaAttack Type=MAD2026.02 | 175.13 | 0.38 | |
| PPOL-vanillaAttack Type=Natural2026.02 | 174.84 | 0.18 | |
| PPOL-vanillaAttack Type=Random2026.02 | 174.63 | 0.22 | |
| TCRL-PPOLAttack Type=Natural2026.02 | 160.51 | 0 | |
| TCRL-PPOLAttack Type=MAD2026.02 | 159.06 | 0 | |
| PPOL-randomAttack Type=MC2026.02 | 158.24 | 40.7 | |
| PPOL-randomAttack Type=Natural2026.02 | 156.29 | 0.48 | |
| ADV-PPOL(MC)Attack Type=Natural2026.02 | 153.5 | 3.7 | |
| PPOL-randomAttack Type=Random2026.02 | 151.39 | 0.38 | |
| PPOL-vanillaAttack Type=Worst-TC2026.02 | 151.28 | 2.72 | |
| PPOL-randomAttack Type=Worst-TC2026.02 | 150.39 | 38.42 | |
| PPOL-randomAttack Type=MAD2026.02 | 148.34 | 1.38 | |
| ADV-PPOL(MC)Attack Type=Worst-TC2026.02 | 115.59 | 2.8 | |
| ADV-PPOL(MC)Attack Type=MC2026.02 | 114.44 | 2.74 | |
| ADV-PPOL(MC)Attack Type=Random2026.02 | 112.92 | 3.28 | |
| ADV-PPOL(MC)Attack Type=MAD2026.02 | 109.5 | 5.38 |