基于3D打印的机械臂数字孪生教学系统开发

Development of a Robotic Arm Digital Twin Teaching System Based on 3D Printing

摘要: 在制造业数字化转型和智能制造课程建设背景下，机械臂作为智能制造核心设备，其教学实践面临实体设备成本高、维护难、抽象机理难以直观呈现等问题。本文基于数字孪生、增强现实(AR)、机器学习等技术，开发了一套机械臂虚实交互实验教学系统。系统以四自由度机械臂为载体，采用SolidWorks完成三维建模与模型轻量化，基于Unity构建虚拟仿真与交互界面，结合STM32主控、HTS-20L总线舵机、HC-05蓝牙通信和串口协议，实现虚拟模型、移动端控制界面与实体机械臂之间的双向同步。在教学设计上，系统围绕结构认知、运动学理解、控制调试和工程反思四类目标，形成“结构认知–虚拟预演–实体验证–数据反馈”的实践教学流程，将D-H运动学建模、正逆解计算、舵机控制和故障保护等内容转化为可观察、可操作、可复盘的学习任务。平台测试结果表明，系统虚实同步延时为185 ms，角度同步误差为1.5˚，舵机定位精度为±0.1˚，故障保护响应时间小于50 ms，能够满足机械臂基础实验教学需求。教学应用分析表明，该系统可支持学生按照“观察模型–预测运动–虚拟调参–实体验证–误差分析”的路径完成实践任务，对降低实体设备依赖和课堂操作风险具有一定应用价值，可为智能制造相关课程的虚实融合实践教学提供参考。

Abstract: Against the background of manufacturing digital transformation and smart manufacturing curriculum development, robotic arms, as core equipment in intelligent manufacturing, face several teaching challenges, including high equipment costs, difficult maintenance, and limited visualization of abstract mechanisms. This paper describes the development of a virtual-physical interactive experimental teaching system for robotic arms, based on technologies such as digital twins, augmented reality (AR) and machine learning. The system utilizes a four-degree-of-freedom robotic arm as its platform. Three-dimensional modelling and model lightweighting were carried out using SolidWorks, whilst the virtual simulation and interactive interface were built using Unity. By integrating an STM32 microcontroller, HTS-20L bus servos, HC-05 Bluetooth communication and serial communication protocols, the system achieves bidirectional synchronisation between the virtual model, the mobile control interface and the physical robotic arm. In terms of instructional design, the system centres on four learning objectives: structural cognition, understanding of kinematics, control debugging and engineering reflection. It establishes a practical teaching workflow of “structural cognition-virtual simulation-physical verification-data feedback”, transforming concepts such as D-H kinematic modelling, forward and inverse solution calculations, servo control and fault protection into observable, operable and reviewable learning tasks. Platform testing results indicate that the system’s virtual-physical synchronisation latency is 185 ms, the angular synchronisation error is 1.5˚, the servo positioning accuracy is ± 0.1˚, and the fault protection response time is less than 50 ms, thereby meeting the requirements for basic robotic arm experimental teaching. An analysis of its educational application indicates that the system enables students to complete practical tasks by following the pathway of “observing the model-predicting motion-virtual parameter tuning-physical verification-error analysis”. It offers practical value in reducing reliance on physical equipment and mitigating operational risks in the classroom, and can serve as a reference for virtual-physical integrated practical teaching in courses related to intelligent manufacturing.