LLM + Reinforcement Learning for Multi-UAV Collaborative Planning
Proposed both coupled and hierarchical planning models for collaborative decision-making and resource scheduling in multi-UAV systems.
LLMReinforcement LearningMulti-Agent SystemsOptimization
Context
This master's thesis targets high-quality and real-time multi-UAV collaborative planning under complex constraints.
Implementation
- Studied collaborative decision-making and resource scheduling with multi-agent systems.
- Proposed two planning paradigms for different scenarios:
- integrated coupled optimization
- hierarchical decoupled planning
Outcome
- Improved global optimization quality under complex mission constraints.
- Maintained practical planning latency for real-time deployment settings.