Decision Support for Submarine Torpedo Countermeasures using Hierarchical Reinforcement Learning
- Hierarchical Multi-Agent Reinforcement Learning Architecture: A high-level agent formulates strategy by selecting optimal low-level agents, which are specialized to execute evasive manoeuvres for different threat scenarios.
- High-Fidelity Underwater Simulation Environment:
- Validated performance in a realistic simulation featuring 6-DOF dynamics and physics-based acoustic models.
- Intelligent Decision Support : Provides commanders with actionable tactical plans (e.g., countermeasures, evasion courses) for complex threat situations.
