Digital Twin-Based Mathematical and Deep Learning Framework for Sustainable Employee Behavior Management in Industrial IoT Environments

Abstract

Energy consumption, safety, productivity, and resource utilization are affected by operational behavior in the Industrial Internet of Things (IIoT) and sustainable cooperation between humans and machines requires human–machine coordination. This study introduces a novel Digital Twin Based Mathematical and Deep Learning Framework (DT-MDLF) that combines five parameters – equipment utilization, energy consumption, safety compliance, task completion efficiency, and idle operation ratio. The mathematical sustainability predictions are achieved by merging a composite mathematical sustainability index with a lightweight deep-learning model to predict the operational behaviour states and to guide sustainability-oriented interventions. The performance of the framework is compared to a threshold-based and standalone deep-learning strategies through a numerical evaluation performed on 100 employee–workstation Digital Twins during 30 days. The proposed DT-MDLF resulted in 22.6% reduction in energy consumption, 27.4% reduction in using idle operations, 96.3% compliance with safety rules, 93.1% task completion efficiency and reduction of sustainability cost from 0.68 to 0.42. The results illustrate the viability of the framework for human–machine coordination in such an IIoT environment, for resource efficiency, safe operation, and sustainability.