[论文解读] Towards the Human Digital Twin: Definition and Design -- A survey

Martin Wolfgang Lauer-Schmaltz, Philip Cash+2 more|arXiv (Cornell University)|Feb 3, 2024

Digital Transformation in Industry被引用 9

一句话总结

本综述定义了 Human Digital Twins (HDTs)，将其分为四种类型，并识别跨领域的设计挑战与考虑因素，以指导未来的 HDT 发展。

ABSTRACT

Human Digital Twins (HDTs) are a fast-emerging technology with significant potential in fields ranging from healthcare to sports. HDTs extend the traditional understanding of Digital Twins by representing humans as the underlying physical entity. This has introduced several significant challenges, including ambiguity in the definition of HDTs and a lack of guidance for their design. This survey brings together the recent advances in the field of HDTs to guide future developers by proposing a first cross-domain definition of HDTs based on their characteristics, as well as eleven key design considerations that emerge from the associated challenges.

研究动机与目标

• 在跨域范围内澄清对 Human Digital Twins (HDTs) 的定义，超越单一用例。
• 识别 HDTs 的类别并将其与传统 Digital Twins (DTs) 区分开。
• 突显 HDTs 中新兴的技术、感知相关以及伦理挑战。
• 提出跨领域 HDT 发展的十一项关键设计考量。

提出的方法

• 对截至 2021 年的 Scopus、Web of Science 和 PubMed 的与 HDT 相关的论文进行系统性文献综述（PRISMA）。
• 将 HDTs 分类为四类：human-in-the-loop、physiological、mechanistic 和 cognitive。
• 综合来自所评审文献的跨领域特征、挑战和设计含义。

Figure 1: Comparison of the three digital concepts for digitally representing a physical entity. The concepts differ in the development of the connecting data thread.

实验结果

研究问题

• RQ1 在一般的跨域背景下，HDTs 如何被定义和概念化？
• RQ2 文献中出现了哪些 HDT 类型？它们与传统 DTs 有何不同？
• RQ3 HDTs 与传统 DTs 差异带来哪些显著挑战？
• RQ4 由这些挑战产生的 HDT 设计考量是什么？

主要发现

• HDTs 通过将认知、生理和行为等人因因素纳入其中，扩展了 DTs，带来了新的设计考量。
• 识别出四类 HDT：human-in-the-loop、physiological、mechanistic 和 cognitive HDTs。
• HDTs 面临由于人类复杂性和动态性所带来的技术概念化挑战，以及感知、接受和伦理方面的挑战。
• 数据质量、决策过程的透明度以及用户界面设计对 HDTs 的信任与接受度具有关键影响。
• 在 HDT 应用之间缺乏统一标准，增加了数据交换与互操作性的复杂性。

Figure 2: Typical Digital Twin framework in conventional application domains. The physical object is monitored using sensors, a continuously updated digital representation is stored using a database, simulations based on Artificial Intelligence are performed and corresponding feedback is provided to