Digital Twin with Embedded Energy Consumption Prediction for Collaborative Robotic Manipulators

Abstract

Collaborative robotic manipulators have become core equipment in intelligent manufacturing, playing an indispensable role in industrial scenarios such as assembly and handling. Driven by the development of Industry 4.0, digital twin technology has become a key supporting technology for process control in intelligent manufacturing. Meanwhile, against the backdrop of global low-carbon transition, energy efficiency has also emerged as an increasingly important consideration in the operation of robotic systems. However, most existing digital twin framework for collaborative robotic manipulators focus primarily on geometric modeling, kinematic synchronization, condition monitoring and trajectory planning, while the integration of energy consumption prediction into the digital twin framework remains largely underexplored, restricting their potential in supporting energy-efficient manufacturing. To fill this research gap, this paper constructs a four-layer digital twin framework for six-degree-of-freedom collaborative manipulators, and embeds an optimized Gradient Boosting Regression Tree (GBRT) data-driven energy consumption prediction model. Experimental results demonstrate that the proposed framework delivers stable and reliable energy consumption prediction capability, enables efficient pre-operation energy assessment, and provides strong technical support for the energy-efficient and stable operation of collaborative robotic manipulators.