Intelligent Anti-Jamming Methods in Complex Electromagnetic Environments of Power Stations

In complex electromagnetic environments, enhancing the anti-jamming capability of wireless sensor networks (WSNs) around power equipment is crucial for ensuring reliable operation. Power equipment generates transient electromagnetic fields and high-frequency transients, leading to conducted and radiated interference, exacerbated by network self-interference. This paper proposes an intelligent anti-jamming system featuring an "Environment-Perception-Agent-Evaluation" (EPAE) closed-loop architecture, utilizing deep reinforcement learning for cognitive reinforcement. The framework integrates: 1) supervised and unsupervised learning for interference recognition, 2) deep reinforcement learning for decision-making, and 3) adaptive array antenna configuration for multi-domain optimization. The EPAE architecture enables dynamic environment awareness, intelligent strategy evolution, and performance evaluation. Simulation results demonstrate that, in dynamically changing real-time interference environments, the proposed method achieves a 32% improvement in interference suppression compared to conventional methods, offering a novel solution for WSNs in power infrastructure.