Intelligent Anti-Jamming Decision-Making Technology Based on Knowledge Graph and DQN.

Area of Science:

• Artificial Intelligence
• Communications Engineering
• Electromagnetic Compatibility

Background:

• Intelligent anti-jamming communications face limitations in reinforcement learning convergence speed and knowledge graph dynamic interaction.
• Complex, time-varying electromagnetic environments pose challenges for existing anti-jamming methods.

Purpose of the Study:

• To propose a novel two-stage intelligent decision-making framework to overcome limitations in current anti-jamming communication methods.
• To enhance real-time responsiveness, model evolution, and self-adaptation in dynamic electromagnetic environments.

Main Methods:

• Constructed an anti-jamming knowledge graph repository for rapid, efficient reasoning and real-time decision-making.
• Introduced a hierarchical reinforcement learning architecture for continuous environmental interaction and model adaptation.
• Simplified multidimensional parameter spaces into two-dimensional scenarios to reduce computational complexity and accelerate convergence.

Main Results:

• Achieved a 4.2% increase in anti-jamming decision success rate compared to state-of-the-art methods.
• Demonstrated a 104.8% improvement in transmission rate.
• Validated superior anti-jamming performance and learning efficiency in dynamic electromagnetic environments.

Conclusions:

• The proposed two-stage framework effectively addresses limitations in existing intelligent anti-jamming communication methods.
• The framework offers practical effectiveness and superior performance in dynamic electromagnetic environments, improving both success rate and learning efficiency.