Autonomous or self-driving vehicles are the future of humans’ go-to commute mode. They use advanced algorithms and sensors without human interaction to navigate their surroundings. However, as vehicles evolve more autonomously due to the lack of human interference, technology reliance makes them vulnerable, leading to increased cyber-attacks. Autonomous vehicles can be susceptible to attack through various means, but the Controller Area Network (CAN) is dangerous as it transmits signals to the Electronic Control Unit (ECU). It creates a lack of reliability in the communication of microcontrollers and devices, making the vehicle prone to accidents. The paper utilizes the Udacity Self-Driving car dataset to replicate a Sybil Attack on the vehicle’s CAN bus. The paper discusses the research methodology, which includes pre-processing the torque and steering angle values, inserting the injected data rows, creating a time-series-based model, and evaluating multiple machine learning models to identify the incorrect values injected into the vehicle's CAN bus. The SVM model outperformed other models and previous research with the highest accuracy of 96.93%. The proposed work leverages the results of various machine learning models to detect cyber-attacks on the CAN system of autonomous vehicles, eventually increasing the overall cyber-related security of vehicles.