A Hybrid Solution to the Human/Machine Inspection Systems Design Problem in Manufacturing and Service Industries

9800817 Gramopadhye这项研究专门解决与人机系统设计有关的问题，因为它适用于检查系统设计。作为研究的一部分，将开发三种不同类型的检查系统：人工、自动化和混合(人+机器)检查。在此步骤之后，将进行对照研究，评估各种实验因素(例如，缺陷数量、缺陷类型、缺陷对比度等)的检查系统的性能。对检查任务的绩效将使用速度、精度、信号检测理论措施和主观评估措施进行衡量。随着不同检查系统的发展和对这三个系统进行的后续研究，将对人工、自动和混合检查系统的相对优点进行评价。此外，本研究还将研究混合检测系统中动态功能分配的相关问题。作为这一步骤的一部分，将考虑各种人机通信替代方案。具体地说，将考虑明确和基于模型的沟通。最后，它将导致开发检查系统中人机功能分配的框架，作为建立最佳系统性能的一步。预计研究结果将有助于更好地了解与人/机检查系统设计有关的问题。这将为系统设计人员提供最大限度地利用人和机器各自的优势来设计最佳性能的检测系统，最终导致检测性能的提高，从而提高质量。这项研究的一般性将有助于将结果不仅应用于制造环境中的各种检查任务，例如印刷电路板检查等，而且还应用于具有重要视觉搜索和决策组件的其他任务(例如，实时军事目标获取、显示监视)。

9800817Gramopadhye The research addresses issues related to human-machine systems design specifically as it applies to inspection system design. As part of the research three different types of inspection systems will be developed: the human, the automated and the hybrid (human + machine) inspection. Following this step, controlled studies will be conducted which will evaluate the performance of the inspection systems for various experimental factors (e.g., number of defects, defect types, defect contrast, etc.). Performance on the inspection task will be measured using speed, accuracy, signal detection theory measures, and subjective evaluation measures. The development of the different inspection systems and the follow up studies conducted on the three systems will lead to an evaluation of the relative merits of the human, automated and hybrid inspection systems. Moreover, the research will also investigate issues related to dynamic function allocation in hybrid inspection systems. As part of this step, various human-machine communication alternatives will be considered. Specifically, explicit and model based communication will be considered. Finally, it will lead to the development of a framework for human-machine function allocation in an inspection system as a step toward establishing the best system performance. It is anticipated that the results of the research will lead to a better understanding of the issues related to human/machine inspection system design. This will provide system designers to make the best use of the respective advantages of humans and machines in designing the best performance inspection systems, ultimately leading to improved inspection performance, and therefore, improved quality. The generic nature of this research will facilitate in applying the results not only to a wide variety of inspection tasks within the manufacturing environment such as printed circuit board inspection, etc., but also to other tasks with significant visual search and decision making components (e.g., real-time military target acquisition, display monitoring).