A Self-Adaptive Digital Twin Architecture to Automate Greenhouse Management

Abstract

Digital Twins are an emerging technology that is increasingly adopted for industrial applications across different domains. Digital Twins can leverage self-adaptive techniques to provide interoperability for cross-domain applications. This interoperability can come with challenges related to the different digital models applied, and with different scenarios in which distributed architectures are in place. In this setting, we propose a self-adaptive digital twin architecture to support reliable automated management for a category of physical twins. We apply declarative dynamic reclassification to capture the appropriate state, and adaptive control through strategies to improve the resilience of the system. The proposed approach is validated in both a simulated environment and a physical environment to compare the efficacy and reliability of the architecture.