Vývoj autonomního řízení vozidla v simulovaném prostředí

Abstract

This master's thesis focuses on the development of autonomous driving systems within the CARLA simulation environment, addressing SAE automation levels and sensory perception. A modular Python architecture integrates assistance systems for longitudinal and lateral control, specifically adaptive cruise control and lane keeping assist. The implementation utilizes radar-semantic data fusion and semantic image analysis, further enhanced by YOLO-based object detection. Finally, the project incorporates custom map and vehicle models alongside a conceptual design for CAN bus communication.