Towards Model-driven Verification of Robot Control Code using Abstract Syntax Trees in Production Systems Engineering

Context. In Production Systems, software components are often tightly connected to defined hardware device types like robots. Different types of robots, even from the same vendor, often use vendor-specific programming languages. Therefore, the exchange of devices or device types, e.g., during system evolution, is challenging and needs new or adapted control software and repeated verification and validation process steps, even if the software behavior remains unchanged. Models aim at supporting these verification and validation tasks during system evolution. Objective. This position paper aims at providing a verification and validation process approach with models for supporting automation systems maintenance and evolution processes. For evaluation purposes, we report on a feasibility study with a focus on two selected robot types in the context of Production Systems Engineering (PSE). Method. We use the Abstract Syntax Tree (AST) concept as a foundation for generating models as the basis for human-based verification and validation. Based on two generated AST variants, related to old and new software control code, human experts can compare the behavior of the expected system to verify and validate the code. Results. First results showed the feasibility of the AST concept to support human-based verification and validation in the context of PSE maintenance projects on a structural level. Conclusion. Although the human-based verification and validation process is feasible and promising on a structural level, the complexity of AST for large-scale models needs to be addressed by tool support to overcome complexity levels of the production system and limitations of human-based verification and validation.