Research on Fusion Detection of FDI Attacks in WSNs with Privacy Preservation Under Communication Constraints

This paper investigates the issue of data security and false data injection (FDI) attack detection and warning speed in wireless sensor networks under bandwidth constraints. Firstly, for non-malicious and FDI attacks, a system model with privacy protection is established to ensure data security and integrity while reducing the covertness of FDI attacks for detection. Secondly, a privacy-protected residual generator is designed to capture the actual trajectory of FDI attacks, effectively identifying and distinguishing between genuine and false data. Introduce multiple quantizers with limited levels to meet communication bandwidth requirements, a convex optimization problem is formulated to design a privacy-protected FDI distributed fusion detection criterion. Subsequently, enhancing the resilience of residual signals to interference and obtain more accurate thresholds, thereby improving detection warning speed. Finally, the effectiveness of the proposed method is validated through simulation results of unmanned surface vessel (USV) navigation at maritime.