Digital Twin-Based Control and Monitoring of Industrial Robotic Arms Using Cloud CAD Platforms

Abstract

This paper presents the design, implementation, and validation of a low-cost Digital Twin framework for controlling and monitoring industrial robotic arms through a cloud-based CAD platform. The system couples a physical prototype—driven by a Raspberry Pi 5, a DRV8825 stepper driver, and an AS5600 magnetic encoder with its virtual counterpart in Onshape, enabling bidirectional synchronization of commands and feedback in near real time. A lightweight Flask web interface provides local control (slider/inputs, start–stop, emergency stop), while RESTful API calls (GET/POST) align CAD mate values with sensor-derived joint states. The closed-loop architecture supports trajectory testing, performance evaluation, and early fault detection prior to execution, reducing risks and downtime. Experimental evaluation shows an average synchronization delay of ~200 ms, mean angular accuracy of ±0.3° at 1/8 microstepping, and response times of ~0.8 s in Local mode versus 1.0–1.5 s via Onshape due to API overhead. Identified challenges encoder calibration with a 3:1 gearbox, API latency, and safety under abrupt stops were mitigated through software mapping, batched updates, and a mushroom-type emergency stop. The framework demonstrates scalability for automated sorting, education, and rapid prototyping, and outlines extensions toward multi-joint systems, richer sensing, collaborative robotics, and predictive maintenance.