This thesis explores the development and effectiveness of an Augmented Reality (AR) system, employing Embodied Augmented Reality Agent (EARA) to facilitate advanced manual machine skill transfer within the industrial workforce. This innovative approach integrates AR technologies with machine perception and artificial intelligence to transcend traditional training methodologies, offering an immersive and interactive training environment.

Central to this research is the design, development, implementation and deployment of an AR system that leverages avatars as instructional medium. These digitally-rendered agents are designed to deliver training sessions in complex machine operations, providing a learning experience that is not only enriched by contextual relevance but also improved by spatial integration. The AR system was constructed using the Microsoft HoloLens 2 (HL2), Unity 3D, Blender 4.0 and underwent a through evaluation involving experts in AR and end user groups to assess its functionality and effectiveness.

The evaluation phases revealed that the AR system enhances the ease of learning and efficiency of learning with which manual tasks are performed. It is demonstrated that the integration of EARA in training results in a reduction in the effort required for training and the improvement of operational efficiency. These findings endorse the effectiveness of AR in improving task execution within industrial settings, highlighting its potential to outperform conventional training techniques.

The findings of this thesis confirm the beneficial potential of AR in industrial training contexts. The AR system developed during this research exemplifies a user friendly and ef-fective system for manual machine task skill training, potentially improving training pro-cesses across diverse sectors. Looking forward, further research is recommended to explore wider applications and enhance the system integration with advanced machine learning models, thus broadening its training and practical implications. This work sets a precedent for future research in the domain of augmented reality applied to workforce training, suggesting a paradigm shift towards more interactive and effective learning frameworks.