A Digital Mentor Apprentice Model for Generative AI-Enabled Practical Training in Rail Transit Smart Operation and Maintenance

DOI: https://doi.org/10.62517/jhve.202616406

Author(s)

Yingge Li*, Hongmei Liu

Affiliation(s)

Guangzhou Railway Polytechnic, Guangzhou, Guangdong, China *Corresponding Author

Abstract

Practical training for the intelligent operation and maintenance of rail transit systems has long faced significant challenges, including high equipment costs, difficulty in replicating faults, a shortage of expert mentors, the inability to standardize experience, and a lack of process visibility. Meanwhile, students' use of general-purpose artificial intelligence tools often suffers from issues such as uncontrolled generalization and a disconnect from instructional objectives. To address these dilemmas, this paper proposes and systematically explores the Digital Mentor Apprentice (DMA) educational model. Leveraging generative AI and multi-modal sensing technologies to create a digital mentor, this model provides each student with one-on-one, end-to-end, and traceable mentorship within a training environment that integrates virtual and physical elements, all while adhering to real-world job standards. The model comprises three key components, digital mentor, virtual-physical integrated training environment, and skill benchmarks based on job standards, and operates through a perception-diagnosis-feedback-progression closed loop driven by the dual engines of multi-modal sensing and generative AI. This paper elucidates the model's conceptual essence, core components, operational mechanisms, and implementation pathways, while also designing a representative instructional scenario focused on fault diagnosis for overhead contact line equipment. The DMA model aim to overcome bottlenecks inherent in modern apprenticeship systems, such as the scarcity of mentors and training scenarios and the lack of process visibility, thereby offering a practical implementation path for AI+major curriculum reform in vocational education.

Keywords

Digital Mentor Apprentice Model; Generative AI; Smart Operation and Maintenance in Rail Transit; Practical Teaching Reform

References

[1]Ma Jun, Ma Junqi. Connotation, constituent elements, and cultivation strategies of digital skills for vocational college students from the perspective of digital labor. Journal of Southwest University (Social Sciences Edition), 2024, 50(1):224-236. [2]Cai Yujing, Gao Fan, Meng Yukun, et al. Design and key technology research of intelligent operation and maintenance systems for urban rail transit equipment. Railway Computer Application, 2023, 32(7):79-83. [3]Gu Xiaoqing, Hu Yiling, Hao Xiangjun. Re-examining artificial intelligence and the future development of education amidst the ChatGPT craze. Academic Digest of Liberal Arts in Higher Education Institutions, 2023, 40(5):152-153. [4]Han Xibin, Li Mixue, Guo Wenxin. New breakthroughs in empowering vocational education through digital strategies: New progress in research and practice of vocational education digitalization in 2024. China Vocational and Technical Education, 2025(2):39-48. [5]Wang Youmei, Wang Dan, Wang Haijie, et al. The deskilling crisis and countermeasures regarding the educational application of generative AI: An analysis based on the framework of transformative reverse thinking. Open Education Research, 2025, 31(4):97-108. [6]An Hong, Lu Fali. Artificial intelligence and vocational education reform: Research review and potential issues. Teaching in Forest Regions, 2023(10):37-42. [7]Han Xibin, Shi Wanruo. Analysis of dilemmas and resolution strategies for virtual simulation training in vocational education. e-Education Research, 2025(9). [8]Lu Yu, Tang Xiaoyu. Morphological hierarchies and progression paths of classroom teaching empowered by generative AI. e-Education Research, 2025,46(6). [9]Liu Bangqi, Nie Xiaolin, Wang Yafei, Yuan Tingting, Zhao Ziqi, and Zhang Guoqiang. Generative AI Empowering Education: Technical Framework, Application Scenarios, and Value. China Educational Technology, 2025(3). [10]Li Yingge, Guo Runhao, Liu Hongmei, et al. A computer vision-enabled closed-loop framework for adaptive practical-skill instruction in higher vocational education: a rail transit maintenance case. Occupation and Professional Education, 2026, 3(6).