Bayesian Transformer-Based Adaptive Identity Authentication for Power Inspection Terminals

Reliable identity authentication of power inspection terminals is vital for power system security. This paper presents a multidimensional authentication method fusing static (e.g., hardware IDs, software configs) and dynamic features (e.g., network params, control parameters). A high-discriminative device fingerprint model is built by analyzing these features. To overcome the limitation of traditional static methods in handling real-time status changes, a dynamic feature analysis mechanism is introduced for real-time operational state perception. Experiments show the method achieves 100% authentication accuracy with an average response time of 235ms. Compared to existing schemes, it offers stronger state adaptability and security, providing an effective solution for terminal authentication. The results contribute to improving power system equipment security management.