How to Train for Industry 5.0? A Comparative Study of Training Strategies for Novice Operators of Collaborative Robots

Abstract

As Industry 5.0 emphasizes a human-centric manufacturing paradigm, effective training of operators for Human-Robot Collaboration is increasingly important. This study compares three pedagogically distinct training strategies, namely Imitation Learning, Cognitive Apprenticeship, and Self-Regulated Learning, to evaluate their suitability for on-boarding novice collaborative robot operators. Novice participants performed a standardized pick-and-place task, assessed through objective performance metrics (effectiveness, efficiency, intervention rate) and subjective experience measures (user trust, cognitive workload). Statistical analysis revealed that Cognitive Apprenticeship significantly outperformed Imitation Learning in reducing both cognitive workload and intervention rate, while the two approaches yielded comparable outcomes on the remaining measures. In contrast, Self-Regulated Learning performed consistently worse across nearly all metrics. These findings indicate that both Cognitive Apprenticeship and Imitation Learning can facilitate novice operator training to varying degrees, whereas Self-Regulated Learning is generally less effective. By systematically evaluating these strategies, this work provides critical insights into the design of adaptable, human-centered training programs for collaborative robots that balance performance with user experience in Industry 5.0 environments. Note to Practitioners—Flexible production often falls short because shop floor workers receive limited or unstructured training, with heavy reliance on self-guidance from written Standard Operating Procedures (SOPs). This study evaluates three distinct training frameworks to guide how different approaches support novice operators in learning to work effectively with collaborative robots. Cognitive Apprenticeship proves the most effective overall, helping novices develop competence and confidence by building a strong foundation of both skill and understanding. Imitation Learning is efficient for teaching specific tasks or standard procedures but does not cultivate the deeper understanding needed to handle unexpected situations. In contrast, Self-Regulated Learning is better suited for advanced users who can troubleshoot and adapt programs independently; it is demanding and generally unsuitable for novices. Overall, even relatively brief training can markedly improve novice capabilities.