Model driven software engineering for IoT edge devices : Leveraging metamodels and NS expansion for software on IoT edge devices in logistical applications.

Internet of Things (IoT) offers significant opportunities to industries, especially in logistics. However, IoT’s heterogeneous character makes software development challenging. Software development typically lacks evolvability, implying the necessity of large quantities of resources to maintain and change software systems over time, critical in IoT due to its many nodes and dynamic environment. This thesis, in collaboration with the Belgian software company NSX, aims to address this challenge by developing a model-driven platform for code generation based on NSX’s evolvable code generators. The developed model-driven platform supports the generation of evolvable software for IoT edge devices, with an extra focus on applications in logistics. Action research is utilized to have a cyclical approach to the development of the model-driven platform, including a literature review. The methodology is further validated and extended through the use of Hevner et al.’s (2004) design science framework. The platform is built upon NSX’s meta-metamodel, Micro Radiant, and includes a metamodel for IoT edge devices and corresponding models. The platform allows for the generation of application code through code templates. The templates incorporate Normalized Systems Theory to ensure evolvability and sustainability. The platform implements a variety of protocols and networks to achieve abstraction from the heterogeneity. This abstraction allows for the generation of a variety of IoT edge applications, including in logistics. The developed regenerative meta-circular loop, allows extensibility and adaptability of the model-driven platform. The integration of evolvability, extensibility and adaptability lays a foundation for sustainable IoT edge software development, addressing gaps in research and industry.