Congestion Attack Detection in Intelligent Traffic Signal System: Combining Empirical and Analytical Methods

The intelligent traffic signal (I-SIG) system aims to perform automatic and optimal control based on situation awareness by leveraging connected vehicle (CV) technology. However, the current algorithm is highly vulnerable CV data spoofing attacks. These vulnerabilities can be exploited create congestion in an intersection even trigger a cascade failure network. To avoid this issue, timely accurate attack detection identification are essential. This work proposes approach combining empirical prediction analytical verification. First, we collect range of images that correspond specific snapshots which potential Based these images, improved generative adversarial network trained predict whether forthcoming will cause with high probability. Meanwhile, define group flow features. After exploring features conducting thorough analysis, TGRU (tree-regularized gated recurrent unit)-based proposed verify occurs. When find possible probability subsequent flows also prove congestion, say there attack. Thus, realize integrating Extensive experiments demonstrate our performs well accuracy timeliness.