An Efficient Intrusion Detection System in IoV Using Improved Random Forest Model

Modern cars use a hierarchical system of sensors, controlling devices, and controllers, linked via various intra-vehicle systems, to regulate and monitor the vehicle's status.Researchers have confined numerous academic papers on intrusion detection in the Internet of Things (IoT), employing data mining and machine learning (ML) techniques to secure autonomous vehicles and detect potential attacks.To identify malicious attacks on the Internet of Vehicles (IoV), however, a competent and accurate method is required.This paper presents a model for cyber-attack detection in IoV that employs tree-based ML methods, an Improved Random Forest Classifier (IRFC), and Extra Tree (ET).We build the proposed model using Improved Random Forest (IRF) and ensemble learning techniques.The proposed IRF model employs optimized feature selection and tuning strategies to enhance intrusion sensitivity and decrease false positive rates.We evaluate the proposed model's performance using the CI-CIDS 2018 dataset.Also, this work focuses mostly on the reduced feature selection and ensemble learning (EL) methods to get a high detection rate while keeping the cost of computing low.The test results show that the proposed method can find DDoS attacks and vehicle intrusions with a 0.99 accuracy rate.