This thesis presents the development of a modular co-simulation framework for testing Advanced Driver Assistance Systems (ADAS) functionalities in autonomous vehicles. The proposed environment integrates high-fidelity vehicle dynamics simulation using VI-CarRealTime with realistic sensor data generation in CARLA, coordinated through ROS2 middleware. The framework enables closed-loop testing of control and perception algorithms implemented in both Simulink and Python, facilitating flexible and scalable experimentation. Key components such as scenario creation, sensor integration, and decision-making modules are described, with a focus on the challenges of real-time communication and system latency. Demonstrations include traffic light recognition and automatic emergency braking in urban scenarios. Limitations and future directions are discussed, including the potential integration of V2X communication using SUMO and optimization of inter-component protocols to enhance real-time performance..