Attack detection and fusion-based security state estimation in multi-sensor networks

This study aims to address the challenges of attack detection and sensor data fusion in multi-sensor stochastic systems where multi-sensor networks are subjected to cyber threats. From the standpoint of minimizing output error power, a detection method is proposed to monitor sensor transmission channels for possible attacks, where the warning system is modeled as a binary hypothesis test. Based on the detection algorithm, the attacked sensor around the fusion center is found, and the attacked sensor signal is filtered by the idea of s-sparse observability. Then, the optimal fusion weight matrix is recalculated by minimizing the error between the fusion center and adjacent sensors and using the Lagrange multiplier method to reconstruct the sensor fusion strategy. Research results confirm that the technique reliably maintains fusion accuracy and robustness in attack scenarios. Through an application in radar target tracking, the secure state estimation results illustrate the method's practical utility.