A Novel Decentralized Analytical Methodology for Cyber Physical Networks Attack Detection

In many functional implementations of considerable engineering significance, cyber physical solutions have recently been developed where protection and privacy are essential. This led to the recent increase in interest in the development of advanced and emerging technology for anomaly and intrusion detection. The paper suggests a new frame for the distributed blind intrusion detection by modelling sensor measurements as the graph signal and using the statistical features of the graph signal for the detection of intrusion. The graphic similarity matrices is generated using the measured data of the sensors as well as the proximity of the sensors to completely take account of the underlying network structure. The scope of the collected data is modelled on the random field Gaussian Markov and the required precision matrix can be determined by adjusting to a graph called Laplacian matrix. For research statistics, the suggested technique for intrusion detection is based on the modified Bayesian probability ratio test and the closed-form expressions are derived. In the end, the time analysis of the actions of the network is calculated by computing the Bhattacharyya distance at consecutive times among the measurement distributions. Experiments are carried out, evaluated and equate the efficiency of the proposed system to the modern method. The findings indicate a detection value better than that offered by other existing systems via the proposed intrusion detection frame.