As emerging technologies advance, Intelligent Transportation Systems (ITS) are transitioning to Autonomous Transportation Systems (ATS). This paper comprehensively outlines and compares the hierarchical mechanisms of ITS across various countries and introduces a theoretical framework for ATS. The framework defines five key elements and their interrelations, proposing a service-oriented hierarchical logical and physical architecture. To validate the ATS theoretical framework, we conduct a case study on autonomous vehicles (AVs) crossing intersections, comparing the efficiency of cloud computing, fixed edge computing, and mobile edge computing. This study captures the scenario architecture through the collaborative relationship of ATS services and conducts traffic flow dynamic simulation. Results indicate that mobile edge computing demonstrates the best performance, enhancing traffic efficiency by approximately 4.9–7.3%. The theoretical framework provides reference for transportation system architects and engineers. Furthermore, this method can be extended to other traffic scenarios, laying the groundwork for practical applications in future industry advancements.