AI-Supported Reform of a Graduate New Energy Vehicle Technology Course: Integrating Multimodal Learning and Value-Oriented Education

Abstract

The rapid development of new energy vehicles and the growing use of artificial intelligence (AI) in engineering practice require graduate students in vehicle engineering to reason across increasingly coupled battery, electric-drive, power-electronics, thermal-management, and control subsystems. This paper reports a design-based reform of the graduate course New Energy Vehicle Technology, undertaken to address fragmented knowledge organization, limited opportunities for open-ended system inquiry, and weak connections between technical learning and professional responsibility. The redesigned course combines a cross-subsystem knowledge graph, digital-twin and co-simulation activities, generative-AI-assisted feedback with mandatory engineering verification, multimodal representations, and project-based learning. Preliminary evaluation draws on platform logs, project work and scores, end-of-course survey responses, and instructor observations. Two representative projects—battery state-of-health estimation with vehicle-range prediction and multi-objective optimization of vehicle energy management—show how students move from component knowledge to system-level analysis and from algorithmic output to engineering judgment. Descriptive course records show that active knowledge-graph use increased from 1.2 to 4.7 queries per student per week, while more than 40% of students conducted additional multi-scenario analyses beyond the minimum requirement. The implementation also exposed uneven preparation in programming and AI-related methods and substantial computing demands for the digital-twin environment. Because the comparison cohort is historical, the findings are interpreted as preliminary rather than causal. The study contributes a course-level design logic for integrating AI-supported learning, multimodal inquiry, and value-oriented education in graduate vehicle engineering.