Modified Self-Organizing Feature Maps for Detection Abnormal Behaviors of Connected Vehicles

Connected vehicles form a self-organized network without a priori fixed infrastructure. However, due to the lack of centralization, they are vulnerable to security attacks, and in order provide security against malicious attacks, Intrusion Detection Systems (IDSs) are being developed for major protection. In this paper, we propose a new scheme for IDSs based on neural networks, which is the Self Organizing Features Maps (SOFM) specifically. We modified this algorithm to improve the performance in detecting anomalies and spot outliers accurately (MSOFM). The privilege of our scheme is that we have no constraints on our data, and thus no need for preprocessing data. The simulations and results demonstrate the capabilities of our scheme in detecting attacks in various scenarios. Keywords-Vehicle; Security; Anomalies; Detection; Neural Network;Adaptive; MSOM, Fake Messages, Communications.