Clustering Methods for Identification of Attacks in IoT Based Traffic Signal System

The traffic signal system plays an important role in smooth ongoing of traffic. The working of signal system will be based on the amount of traffic coming towards or passing across the junction. There must be some sort of communication needed to let the signal system know about the number of vehicles driving towards the signal point. Whenever there is a communication especially wireless, the chances of an attacker in the middle of communication can be more. To avoid attacks of the kind like same signal lasting for more time or same signal on right and left turns at the same time which might leads to vehicle crashes are to be detected and rectified for better working of the road systems. In this paper, we focus on detecting those attacks using different machine learning concepts and analyzed the results for better understanding of algorithms and their role in detecting attacks. We are applying the models on a real-time dataset and results are analyzed. Finally, the paper results out the best clustering algorithm to detect the attacks in traffic signal system data and models are compared under 4 different parameters.