Metadata association feature ATC data security assessment

Abstract

The assessment of air traffic management (ATM) security is important for maintaining the secure operation of ATM information systems. However, the accuracy of ATM assessment still needs to be improved. To solve this problem, this paper proposes a security assessment method for ATM data based on metadata correlation characteristics. The method consists of two parts: calculating the weight characteristics and optimizing the evaluation model. In the stage of calculating weight features, we extract ATM features from metadata with normalization method to obtain evaluation indicators for weight allocation. Then, the fuzzy Borda method and CRITIC method are used for weight assignment. The variable weight synthesis method is used to dynamically modify the weight, and finally the normalization method is used to achieve dimensionless processing of indicators. In the stage of optimizing the evaluation model, the multi-layer feedforward neural network is used to optimize the weights parameters. Compared with comparison methods, the accuracy of our method reaches up to 97 %, while the accuracy of compared methods fluctuates between 40 % and 80 %. In our method, the safety assessment time is up to maximum 12 s, the confidence level is always above 95 % and the p-value of the assessment results around 0.95. Comparative experimental results show that the proposed method can improve the accuracy of ATC safety assessment, and is of great significance to promote the integrity of ATM safety risk assessment system.