Digital-Twin-Based Real-Time Optimization for a Fractional Order Controller for Industrial Robots

Abstract

Digital twins are applied in smart manufacturing towards a smarter cyber-physical manufacturing system for effective analysis, fault diagnosis, and system optimization of a physical system. In this paper, a framework applying a digital twin to industrial robots is proposed and realizes the real-time monitoring and performance optimization of industrial robots. This framework includes multi-domain modeling, behavioral matching, control optimization, and parameter updating. The properties of the industrial robot are first modeled in a digital environment to realize the strong interactive and all-around 3D visual monitoring. Then, behavioral matching is performed to map the virtual system to the physical system in real time. Furthermore, the control performance of the system is improved by using a fractional order controller based on the improved particle swarm optimization algorithm. This framework is applied to the experimental verification of real-time control optimization on an industrial robot. The time-domain performance is improved in the simulation and experimental results, where the overshoot is promoted at least 42%, the peak time is promoted at least 32%, and the settling time is promoted at least 33%. The simulation and experimental results demonstrate the effectiveness of the proposed framework for a digital twin combined with fractional order control (FOC).