Three-Level Intrusion Detection Model for Wireless Sensor Networks Based on Dynamic Trust Evaluation

In the complex environment of Wireless Sensor Networks (WSNs), various malicious attacks have emerged, among which internal attacks pose particularly severe security risks. These attacks seriously threaten network stability, data transmission reliability, and overall performance. To effectively address this issue and significantly improve intrusion detection speed, accuracy, and resistance to malicious attacks, this research designs a Three-level Intrusion Detection Model based on Dynamic Trust Evaluation (TIDM-DTE). This study conducts a detailed analysis of how different attack types impact node trust and establishes node models for data trust, communication trust, and energy consumption trust by focusing on characteristics such as continuous packet loss and energy consumption changes. By dynamically predicting node trust values using the grey Markov model, the model accurately and sensitively reflects changes in node trust levels during attacks. Additionally, DBSCAN (Density-Based Spatial Clustering of Applications with Noise) data noise monitoring technology is employed to quickly identify attacked nodes, while a trust recovery mechanism restores the trust of temporarily faulty nodes to reduce False Alarm Rate. Simulation results demonstrate that TIDM-DTE achieves high detection rates, fast detection speed, and low False Alarm Rate when identifying various network attacks, including selective forwarding attacks, Sybil attacks, switch attacks, and black hole attacks. TIDM-DTE significantly enhances network security, ensures secure and reliable data transmission, moderately improves network energy efficiency, reduces unnecessary energy consumption, and provides strong support for the stable operation of WSNs. Meanwhile, the research findings offer new ideas and methods for WSN security protection, possessing important theoretical significance and practical application value.