An adaptive AR guidance interface layout optimization approach for human-centered assembly training systems

In industrial assembly, Augmented Reality (AR) technology plays a critical role in reducing training time and enhancing operational efficiency. However, the limited field of view (FOV) of head-mounted displays (HMDs) remains a significant challenge. Current AR interface paradigms, which typically employ fixed-position or manually interactive layouts, often suffer from poor integration with the physical assembly environment. Such conventional approaches often neglect critical user physiological and cognitive factors, resulting in a suboptimal user experience and impeding the broader adoption of AR-assisted assembly systems. To address these limitations, this paper proposes a user-centric adaptive spatial layout mechanism. This system dynamically optimizes the AR interface by monitoring the user’s posture and FOV on demand. Our method integrates ergonomic principles, individual user physical characteristics, and device constraints into a multidimensional optimization algorithm. The algorithm adjusts the interface’s position in terms of depth, vertical, and horizontal dimensions to ensure the AR content remains consistently aligned with the user’s natural line of sight and maintains a comfortable posture within the FOV of the HMD. A prototype system was developed on the Microsoft HoloLens 2 platform for a seated manual assembly task. A comparative user study against fixed-position and manual interactive layouts demonstrated that the proposed adaptive system yields statistically significant improvements in reducing neck and shoulder fatigue, improving adjustment efficiency, and enhancing the overall user experience. The experimental results confirm the effectiveness and innovativeness of the proposed adaptive method, underscoring its potential to advance human-centric AR training systems.