Manual Assembly Augmented Reality Systems Implementation: A Systematic Literature Mapping

Manual assembly operations can benefit from AR-based systems but the industry seems not to follow the idea. A gap between conceptual systems and actual ones on the shop floor in the industry might exist. Recommendations and guidelines for the development of AR-based manual assembly systems can be found, however, the prevalence of such applications in everyday industry remains relatively low. Due to the absence of reviews targeting systems' application and usage, we performed a systematic literature mapping focusing on the implementation of AR-based manual assembly systems in real settings and the adaptations required. This paper reviewed the research on using AR to assist manual assembly from 2012 to 2022 and from 3330 papers a total of 77 were selected. The majority of the proposed systems were applied to a simulated environment and were evaluated by subjects that were not directly related to manual assemblies. It was also found that among the wide variety of tasks related to the presented manual assembly AR systems, the most common was Lego-based assembly. A preference was observed for manual assembly AR systems that display the actual assembly procedure, and provide guidance by presenting instructions within the visual field where the assembly is taking place. Regarding the main aspects of AR systems that have been evaluated to validate the results, we found: assembly time, number of errors, user experience, and cognitive load. Regarding implementing a manual assembly AR system in actual settings, it was found that the system should display only essential orientation to the task and be adaptative to the user's knowledge. Our analysis indicates that current AR-based manual assembly systems are still in the process of identifying the necessary requirements for achieving an efficient and ergonomic system. Therefore, they have yet to reach widespread adoption as a common technology in the industrial setting.