STTR Phase I: Intelligent Instruction Systems using Augmented Reality

这个小型企业技术转移第一阶段项目旨在加强生产线制造培训流程的当前最先进水平，最初的重点是汽车行业。其他技术熟练的雇员和主管以人为本的培训没有利用最新的技术，这些技术可以使培训程序更具自主性，从而在资源有限的时期减轻对培训质量的影响。经验培训结果(即目前无法获得的量化培训反馈)可以帮助计划员优化制造流程。拟议的创新在于创建利用自适应软件机制(即智能软件代理)和增强现实(AR)技术的智能教学系统，从而增强培训技术，同时促进持续的过程标准化、评估、优化和资源管理。由于生产线员工被要求戴护目镜是很常见的，智能代理可以通过类似护目镜的可穿戴计算机(即，AR)来传输他们的指令，这些计算机覆盖在实际的视野中，覆盖着文本和计算机图形。这项研究的一个主要方面是将这些技术插入到实际制造环境中的体系结构、框架和可行性分析。初步分析将针对通用汽车和宝马两家工厂。拟议的技术将能够在培训例程期间对员工进行实时评估，并使软件代理能够根据这些评估自动和主动地加强薄弱领域。所有员工的全面评估模型可以描述特定设施的整个员工队伍的特征。这一总体评估可用于加强持续存在的缺勤问题所需的资源管理。也许最大的创新将是一个框架，允许计划者利用评估并通过重构传统的、可能过时的生产过程来优化制造过程。如果智能代理能够管理和指导任何行业员工的交叉培训，那么一个主要的回报将是资源的高效利用。此外，如共同主张所表达的那样，使用这种试剂将加强以下方面：“过程标准化导致最终产品的可预测性”。如果所有员工都接受了持续执行类似任务的培训，并且智能代理和AR可以执行该标准化，那么质量也将变得可预测和可衡量。汽车以外的制造环境将拥有使其流程标准化的模型、框架、工具和技术。因此，向任何行业的外部客户交付的产品质量都可以提高，同时可以降低生产成本。

This Small Business Technology Transfer Phase I project seeks to enhance the current state of-the-art in production line manufacturing training processes with an initial focus on the automotive industry. Human-directed training by other skilled employees and supervisors does not take advantage of recent technologies that can make training routines more autonomous, thereby mitigating the impact on the quality of training in periods where resources are constrained. Empirical training results (i.e., quantitative training feedback that is not currently available) can assist planners in optimizing manufacturing processes. The proposed innovation lies in the creation of intelligent instruction systems that exploit adaptive software mechanisms (i.e., intelligent software agents) and augmented reality (AR) techniques, thus enhancing training techniques while promoting continuous process standardization, assessment, optimization and resource management. Since it is common that production-line employees are required to wear goggles, intelligent agents could transfer their instruction via goggle-like wearable computers (i.e., AR) that overlay the actual visual field with text and computer graphics. A major aspect of this research is the architecture, framework, and feasibility analysis of the insertion of these technologies into real manufacturing environments. Initial analysis will be with two facilities, General Motors and BMW. The proposed techniques will enable real-time assessment of employees during training routines and enable the software agents to automatically and proactively reinforce weaker areas based on these assessments. An overall assessment model of all employees can characterize the entire workforce for a particular facility. This overall assessment can be used to enhance resource management required by the on-going problem of absenteeism. Probably the greatest innovation would be a framework to allow planners to exploit the assessments and optimize manufacturing processes by refactoring traditional, perhaps obsolete, production processes. If intelligent agents could manage and direct the cross training of employees in any industry, a major pay-off would be the efficient use of resources. In addition, the use of such agents would enhance the following as expressed by the common claim: "Standardization of processes results in the predictability of the final product". If all employees were trained to perform similar tasks consistently and intelligent agents and AR could enforce that standardization, quality would then also become predictable and measurable. Manufacturing environments beyond automotive will have the models, frameworks, tools, and techniques to standardize their processes. As a result, the quality of products delivered to external customers in any industry can be improved, while at the same time production costs can be reduced.