A hybrid machine learning model for intrusion detection in VANET

• Published in: Computing Vol. 104; no. 3; pp. 503 - 531
• Main Authors: Bangui, Hind, Ge, Mouzhi, Buhnova, Barbora
• Format: Journal Article
• Language: English
• Published: Vienna Springer Vienna 01.03.2022; Springer Nature B.V
• Subjects: Accuracy; Artificial Intelligence; Big Data; Collaboration; Communication; Communications traffic; Computer Appl. in Administrative Data Processing; Computer Communication Networks; Computer Science; Data exchange; Efficiency; Information Systems Applications (incl.Internet); Intrusion detection systems; Machine learning; Mobile ad hoc networks; Smart cities; Software Engineering; Special Issue Article; Traffic congestion; Urban environments; Vehicles; 68T01; Intrusion; Machine learning; Classification; VANET; Security; Clustering; Coresets; Random Forest
• ISSN: 0010-485X; 1436-5057
• DOI: 10.1007/s00607-021-01001-0

While Vehicular Ad-hoc Network (VANET) is developed to enable effective vehicle communication and traffic information exchange, VANET is also vulnerable to different security attacks, such as DOS attacks. The usage of an intrusion detection system (IDS) is one possible solution for preventing attacks in VANET. However, dealing with a large amount of vehicular data that keep growing in the urban environment is still an critical challenge for IDSs. This paper, therefore, proposes a new machine learning model to improve the performance of IDSs by using Random Forest and a posterior detection based on coresets to improve the detection accuracy and increase detection efficiency. The experimental results show that the proposed machine learning model can significantly enhance the detection accuracy compared to classical application of machine learning models.