Nowadays, “Autonomous Vehicles (AV)”, also called “ Self-Driving Vehicles (SDV)”, use Artificial Intelligence (AI) techniques/approaches, and provide a large number of benefits ranging from safety (i.e., reducing car accidents and junks) to entertainment functionalities. To deliver these emergent technologies, SDV should communicate with the environment through long-range Vehicle to everything (V2X), Vehicle-to-Infrastructure (V2I) and Vehicle-to-Vehicle (V2V)) and short-range (Near-field communication (NFC), Bluetooth) technologies. Hence, by connecting to the vehicle, wirelessly or physically (OBD-two port, USB), an attacker can exploit the intra-vehicle/in-vehicle network vulnerabilities and have full access/control of its functionalities and services. These attacks can lead to catastrophic consequences ranging from human safety to economic losses. In fact, an in-vehicle network is composed of different Electronic Controlling Unit (ECU) that exchange various messages/data/information through diverse com munication technologies such as the Controller Area Network (CAN), Automo tive Ethernet, FlexRay, Media Oriented Systems Transport (MOST). Through our work, we focus on the CAN technology which lacks important security fea tures such as authentication and encryption, resulting in cyberattacks on the ve hicle such as Denial of Service (DoS), fuzzing, and targeted attacks. Therefore, we aim to develop an intrusion detection system to monitor the CAN traffic and detect abnormal behavior using Machine Learning (ML) techniques, with low false positive and miss detection rates.