Scholarlink

<p>This study presents a novel approach for transferring expert tacit knowledge to construction students through a cross-modal deep learning supported mixed reality training system. Traditional construction training faces challenges in effectively capturing and transmitting the implicit expertise of experienced practitioners. We developed a framework that integrates multimodal data collection from construction experts with cross-modal deep learning algorithms to identify critical knowledge patterns and represent them in an immersive mixed reality environment. Experimental validation with 80 participants demonstrated that students trained using this system achieved 32.4% faster skill acquisition, 22.4% higher accuracy in construction tasks, and significantly better knowledge retention compared to traditional methods. The system particularly excelled in transferring situational awareness and adaptive problem-solving capabilities. Results indicate that cross-modal deep learning approaches can successfully formalize expert tacit knowledge into learnable representations, providing a scalable solution for construction education that addresses industry challenges of workforce development and knowledge preservation. Despite these promising results, the study acknowledges limitations including computational latency challenges during complex interactions and hardware ergonomic constraints during extended training sessions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cross-modal deep learning enhanced mixed reality accelerates construction skill transfer from experts to students

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月2日

查看全文
  • Chenyu Wang,
  • Meng Zhu,
  • Sharifah Akman Syed Zakaria

摘要

This study presents a novel approach for transferring expert tacit knowledge to construction students through a cross-modal deep learning supported mixed reality training system. Traditional construction training faces challenges in effectively capturing and transmitting the implicit expertise of experienced practitioners. We developed a framework that integrates multimodal data collection from construction experts with cross-modal deep learning algorithms to identify critical knowledge patterns and represent them in an immersive mixed reality environment. Experimental validation with 80 participants demonstrated that students trained using this system achieved 32.4% faster skill acquisition, 22.4% higher accuracy in construction tasks, and significantly better knowledge retention compared to traditional methods. The system particularly excelled in transferring situational awareness and adaptive problem-solving capabilities. Results indicate that cross-modal deep learning approaches can successfully formalize expert tacit knowledge into learnable representations, providing a scalable solution for construction education that addresses industry challenges of workforce development and knowledge preservation. Despite these promising results, the study acknowledges limitations including computational latency challenges during complex interactions and hardware ergonomic constraints during extended training sessions.