Non-Functional Requirements Elicitation Based on Domain Knowledge Graph for Automatic Code Generation of Industrial Cyber-Physical Systems

As an important part of Industry 4.0, industrial software development must be highly flexible and reconfigurable to deal with a large number of requirements and random failures during the manufacturing process. Requirements understanding errors that occur in the early stages of software development often cannot be discovered prior to the testing phase. Therefore, the analysis and understanding of requirements are particularly crucial in the software development process. In this paper, a knowledge-driven functional-oriented requirements elicitation and analysis method are proposed to extract both functional and non-functional requirements from design documents. Nonfunctional requirements which are implicit in the documents will be elicited according to the domain knowledge graphs and predefined rules. A domain glossary is also constructed for the elicitation process, especially from unstructured sources. Domain knowledge graphs are modified periodically according to the feedback information to form a closed loop from requirements to code. The feasibility of the proposed method is demonstrated under an automobile welding line.