This paper extends our previous study on kinematics and dynamics control of self-driving cars. The analysis results of the kinematics of an autonomous four-wheel vehicle and an aspect of the vehicle’s perception applying for control of a four-wheeled electric vehicle (Calequela et al in Eur Phys J Spec Top 230:3663–3672 (2021)). In this research, the authors utilize kinematics and dynamics models of the autonomous four-wheel vehicle from previous research and develop a lane detection model for the vehicle control system, set up the environment for simulation, and evaluate performance. The lane detection based on CNN is applied for autonomous controlling of the vehicle. With the video captured from the autonomous vehicle's camera, the lane along with other objects will be recognized, computed, and segmented, thereby determining the safety trajectory and deviation from the autonomous vehicle's intended trajectory. The PID controller relies on the detected trajectory deviations calculated based on the results of the recognition model, thereby adjusting the power supplication to each wheel to adjust the vehicle movement to the desired trajectory. The simulation of the vehicle is conducted using MATLAB-Simulink.