Digital Twin Features for the Intelligent Container

The “Intelligent Container” for remote monitoring of refrigerated transports of fresh fruits already implements typical features of digital twins, including remote sensing and modeling of physical and biological objects. This article asks how the Intelligent Container can be extended to make the best use of digital twin concepts. Existing applications in agricultural science focus on offline simulation models that can predict shelf life and the effects of packaging on cooling but cannot integrate real-time data or correct their current estimates according to those data. This update feature is considered a key component of digital twins. The related challenges and algorithms can be best understood from the viewpoint of systems theory and state-space description. Internal properties of real objects can be either directly measurable, hidden, or unobservable, and implementation of the update process should be adapted accordingly. Using ocean transport and banana processing as an example, this paper demonstrates how models can be made “updateable”, in addition to discussing the necessary steps for linking different sub-models over a standardized platform according to the “publish/subscribe” pattern.