An Autonomous Transportation System Architecture Mapping Relation Generation Method Based on Text Analysis

Information technology and personalized demands are driving the transportation system from passive transportation to active service provision. Autonomous transportation systems (ATSs) are resultantly emerging. A strongly related system architecture can effectively support the planning, construction, and operation of ATS. The current research on the relationship between architectures focuses on the definition of elements and the description of interfaces; however, insufficient research has examined the mapping relationship of the architectural elements. Therefore, to construct a set of architecture systems with explicit mapping relationships, we proposed an information flow (IF) extraction method based on data flow (DF) text to realize automated mapping between logical DF and physical IF. First, to better describe the ATS, we deconstructed the system with service as the basic unit, introduced the structure and constituent elements of the logical and physical architecture, and summarized the mapping principles of the architectural elements. Next, we extracted the DF set based on flow attributes and implemented the vectorization of the DF text. Then, we obtained the similarity of the DFs through text semantic analysis and completed the identification of similar DFs. Finally, to realize the independent construction of IF, we designed a graph-based DF aggregation method to complete the automated mapping of architectural elements. Experimental results showed that the construction method of IF proposed in this article met the requirements of ATS architecture and could effectively support the design and construction of the ATS architecture system.