Research map / Reinforcement learning / Visual and representation RL
Realistic Curriculum Reinforcement Learning for Autonomous and Sustainable Marine Vessel Navigation
Xiaocai Zhang, Zhe Xiao, Maohan Liang, Tao Liu, Haijiang Li, Wenbin Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 2 (46 h) | Accumulated Reward: -708.765 -> 294.148 | Improved |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 1 (21 h) | Accumulated Reward: 139.629 -> 154.018 | Improved |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 1 (21 h) | Accumulated Fuel Consumption: 21.893 -> 18.015 | Improved |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 3 (21 h) | Accumulated Safety Score (lower is safer): 5.901 -> 2.143 | Improved |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 3 (21 h) | Accumulated Fuel Consumption: 19.557 -> 28.618 | Hurt |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 3 (21 h) | Accumulated Reward: 149.950 -> 145.015 | Hurt |
| add curriculum learning to PPO agent (ablation) | CRL · Navigation Instance Case 2 (46 h) | Accumulated Safety Score (lower is safer): 3.274 -> 4.754 | Hurt |
| add curriculum learning (shrinking goal threshold) during training (ablation) | CRL · training reward curve | convergence speed and stability | Improved |
| use PPO-based CRL instead of CL-A2C | vessel navigation agent · Navigation Instance Case 1 (21 h) | Accumulated Reward: -82.034 -> 154.018 | Improved |
| use PPO-based CRL instead of DDPG | vessel navigation agent · Navigation Instance Case 1 (21 h) | Accumulated Fuel Consumption: 33.533 -> 18.015 | Improved |
| use PPO-based CRL instead of CL-ABDDQN | vessel navigation agent · Navigation Instance Cases 1-3 | Accumulated Reward: Case 1: 155.527 vs 154.018; Case 2: 298.026 vs 294.148; Case 3: 155.035 vs 145.015 | Mixed |
| use PPO-based CRL instead of DDPG | vessel navigation agent · Navigation Instance Case 3 (21 h) | Accumulated Fuel Consumption: 9.002 -> 28.618 | Hurt |
Closest papers by meaning
- Sequential Modeling of Complex Marine Navigation: Case Study on a Passenger Vessel (Student Abstract) (AAAI 2024 · similarity 0.87)
- RENEW: Risk- and Energy-Aware Navigation in Dynamic Waterways (AAAI 2026 · similarity 0.86)
- OCEAN-MBRL: Offline Conservative Exploration for Model-Based Offline Reinforcement Learning (AAAI 2024 · similarity 0.83)
- Reducing AUV Energy Consumption Through Dynamic Sensor Directions Switching via Deep Reinforcement Learning (AAAI 2025 · similarity 0.83)
- Dyna-Style Safety Augmented Reinforcement Learning: Staying Safe in the Face of Uncertainty (NeurIPS 2026 · similarity 0.82)
- PWM: Policy Learning with Multi-Task World Models (ICLR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
