Mars Lander Trajectory Generation on RL-25
3.35Cumulative Control CostBPPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BPPOTraining Samples=25, Training Source=S-VAE-25, Generated Trajectories=200, Reward Function=Equation (23)2026.04 | 3.35 | 15 | |
| S-VAEType=Dataset Samples2026.04 | 3.36 | 3.6 | |
| BCTraining Samples=25, Training Source=S-VAE-25, Generated Trajectories=2002026.04 | 3.45 | 15.3 | |
| BCTraining Samples=25, Training Source=RL-25, Generated Trajectories=2002026.04 | 3.48 | 19.4 | |
| RL-25 (Source Data)Type=Dataset Baseline2026.04 | 3.99 | 4.8 | |
| BPPOTraining Samples=25, Training Source=RL-25, Generated Trajectories=200, Reward Function=Equation (23)2026.04 | 4 | 14.4 |