Feature Selection Model-Based Intrusion Detection System for Cyberattacks on the Internet of Vehicles Using Cat and Mouse Optimizer

Abstract

The Internet of Vehicles (IoV) is an environment that is changing quickly. To protect user data and car safety, cyber security is very important. Choosing the right features is important for making intruder detection systems (IDS) more useful and efficient in Internet of Things (IoT) situations. This paper shows a new way to use optimization methods for feature selection-based intruder detection in IoV. Optimization methods are used in the proposed framework for selecting the most important attributes. This makes detection more accurate and reduces the cost of computing. The population-based optimization technique formula is a linked, important, and helpful way to resolve optimization issues. The Cat and Mouse Optimizer (CMO) is a new way to optimize based on how mice and cats naturally act. The proposed CMO acts like cats attacking mice and mice running away to safe places. Popular methods like Kernel Linear Discriminant Analysis (KDA) and the Cuckoo technique are used to compare the CMBO's results. The research showed that the CMO-Bi-LSTM model was very accurate on the CICIDS-2018 and Car hacking datasets, with scores of 99.36% and 99.80%, respectively. Testing and analysis show that the proposed approach successfully finds attacks while reducing false positives. The flexible and expandable nature of the system makes it perfect for securing IoV settings from new hacking attacks.