This tool provides a vision-based control method for autonomous vehicle lane-keeping. This approach aims to provide a reliable alternative solution for localization and path planning-related issues in environments where GPS is unavailable or unreliable, such as in tunnels and urban areas. The proposed method consists of a robust lane detection algorithm to generate a reference path of the vehicle and a model predictive controller (MPC) for tracking the reference path. The lane detector extracts lane markings from image frames to determine ego-lane boundaries, from which the lane center is calculated in the vehicle’s coordinate frame. The MPC uses a kinematic vehicle model to generate the lateral and longitudinal control values necessary for smoothly tracking the reference path. The proposed technique has been implemented and tested on a Lincoln MKZ hybrid vehicle equipped with a computer powered by Intel’s quad-core Xeon processors. Experimental results demonstrate that the proposed vision-based lane detection method performs well under various challenging road conditions, such as shadows, road texture variations, interference from other road signs, and missing lane boundaries. 

Link to the tool: https://github.com/ACCESSLab/vision_control?tab=readme-ov-file 

Related Publication: T. Getahun and A. Karimoddini, "An Integrated Vision-Based Perception and Control for Lane Keeping of Autonomous Vehicles," in IEEE Transactions on Intelligent Transportation Systems, vol. 25, no. 8, pp. 9001-9015, Aug. 2024, doi: 10.1109/TITS.2024.3376516. https://ieeexplore.ieee.org/document/10480914