关于军事行动中医疗疏散的决策支持系统建议书

***传感器***

![](/media/202412//1733390409.0901184.jpeg)

文章

关于军事行动中医疗疏散的决策支持系统建议书t

彼得 ·卢布科夫斯基1号，*，贾洛斯瓦夫 ·克里吉尔![](/media/202412//1733390409.1019304.png)![](/media/202412//1733390409.1095896.png)1![](/media/202412//1733390409.1177793.png)Tadeusz Sondej1，安德泽普。多布罗沃尔斯基1![](/media/202412//1733390409.1274152.png) ，卢卡斯 · 阿皮 利奥内克2,3Znaniecki3和Pawel奥斯卡雷克4

1 军理工大学电子学院。波兰华沙；法律。krygier@wat.edu.pl (J.K.)；塔迪乌斯。sondej@wat.edu.pl (T.S.) ; 安德烈。dobrowolski@wat.edu.pl (A。P.D.)

维尔基大学计算机科学研究所，

2 波兰85-064； 电话。com.pl

Teldat Sp.z o.o.sp.波兰比德戈什茨市，第19、85-650号。com.pl

3 军事医学研究所-国家研究所，萨泽罗128,04-141，波兰华沙。mil.pl

4 通信：piotr。lubkowski@wat.edu.电话 。 : +48-261-837-897

![](/media/202412//1733390409.1381464.png)

![](/media/202412//1733390409.1468298.png)引文：卢布考斯基，P.；Krygier，J. ; SondejT.t；多布罗沃尔斯基，一个 。P.；，l；奥斯卡warek，P。关于军 事行动中医疗疏散的决策支持系统建 议书。传感器2023、23、5144。

https://

doi.org/10.3390/s23115144

学术编辑：Tamura东雄

收稿日期：2023年4月28日

修订日期：2023年5月20日

接受日期：2023年5月26日

出版：2023年5月28日

*

本文由No.项目支持DOB-SZAFIR/09/B/006/01/202 1 ， 由 国家研究中心资助 研发部（NCBR）-波兰。

摘要：军事行动领域是医疗支持面临的一大挑战。使医疗服务部门能够在大规模伤亡情况下迅速作出反应 的一个特别重要的因素是从战场上迅速撤离受伤士兵的能力。为了满足这一要求，一个有效的医疗后送系 统至关重要。本文介绍了军事行动中医疗后送的电子支持决策支持系统的体系结构。该系统也可被警察或 消防部门等其他服务机构使用。该系统满足战术作战伤亡护理程序的要求， 由测量子系统、数据传输子系 统和分析推理子系统组成。该系统基于对选定士兵的生命体征和生物医学信号的持续监测， 自动提出了对 受伤士兵的医疗隔离（医疗分诊）。如果需要，使用总部管理系统为医务人员（急救人员、医务人员、医 疗后送小组）和指挥官可视化了关于分诊的信息。本文描述了该体系结构的所有元素。

关键词：企业架构、决策支持系统、医疗后送、生命体征测量；生物医学信号测量

1.介绍

医疗后送，也简称医疗疏散，是医务人员从战场或事故现场提供的一系列活动。疏散过程 的第一阶段致力于在事故地点对病人挽救生命的医疗护理，然后由医疗人员乘坐紧急医疗服务 车辆或航空医疗直升机提供途中护理。在典型情况下，应急服务人员在收到直接从受伤人员或 事故目击者那里收到的事故信息后作出反应。在事故地点提供急救后， 由救护车（医疗救援人 员）进行医疗运输的决定。如果我们发生了许多不同严重程度的伤亡事故，情况可能会更加复 杂。通常， 由于缺乏资源，医务人员必须决定哪些病人必须首先得到帮助，哪些病人可以等待 。在医学上，这种选择被称为分诊。

![](/media/202412//1733390409.1626225.jpeg)

![](/media/202412//1733390409.1712794.jpeg)

知识共享的条件

归属（CC BY）许可证。

org/licenses/by/ 4.0/).

![](/media/202412//1733390409.1809742.png)

传感器2023、23、5144。https://doi.org/10.3390/s23115144 https://www.mdpi.com/journal/sensors

因此，可以由事故现场的医务人员进行第一次分诊，以决定必须首先疏散到医院和随后可 以疏散的病人的顺序。在战场上，许多受伤士兵在军事行动中突然出现，但在和平时期的事故 中，也需要如此。为了有效支持分诊决定，减少医疗后送时间，医务人员在到达事故现场之前 就应配备有关伤口严重性和病人病情的信息。为此，需要一个有效的医疗后送决策支持系统（D SS-MEDEVAC）。本文的作者提出了这样一个军事用途的系统，其元素可能被其他服务部门，如 警察、消防部门或有组织的救援组织使用。该系统是从零开始开发的，并建议为由指挥和控制 信息系统支持的武装部队提供支持。dss-医疗医疗系统的主要目标是监测士兵在军事行动期间 的健康参数，处理这些参数，并在确定士兵的生命和健康风险时，将建议的军事分类决定提交 给军事人员。同时，关于潜在医疗问题的信息可以传递给指挥官，以支持他们在军事行动期间 的决定。根据医疗的信息，负责战术疏散护理的医务人员能够对战场上的情况做出更快的反应 , 拥有关于建议的分诊和伤亡健康的最新信息。这一信息支持医务人员参与关于医疗后送的决 策过程。此外，这些信息还被传递给医疗后送车辆或直升机，以不断支持疏散人员。

DSS-MEDEVAC系统由健康监测传感器集成了士兵的制服和个人设备，通信模块负责可靠的 医疗数据传输，决策支持模块负责处理从每个士兵的传感器收到的数据和采取初步的决定分类 。这一信息有助于医务人员，他能够在撤离之前和期间就医疗后送和医疗护理作出最终决定。 此外，疏散车辆的医疗人员可以在战场急救或途中护理后改变系统建议的决策。新数据由dss- 医疗服务系统处理，并自动传输到已部署或固定的医疗设施（决策支持中心、疏散点、野战医 院）。dss-医疗治疗系统根据其医疗传感器提供的信息自动填写电子TCCC卡（战术战斗伤亡护 理卡）。该电子卡持续提供当前的患者数据趋势，可以取代目前使用的纸质TCCC卡，该卡附在 受伤士兵身上进行疏散，只有在可能的情况下才能更新。因此，医疗设施的人员可以立即利用 有关伤亡人员的信息， 以准备所需的医疗资源。

本文着重介绍了DSS-MEDEVAC系统架构、系统组件的一般说明以及这些组件的首次实验室 测试。dss-医疗服务系统是一个预先验证的系统，建议在常规军事单位中实施。该系统计划在 波兰语应用

武装部队结束后的现场测试。

我们可以区分我们的系统的以下优点：

![](/media/202412//1733390409.1917396.png)它支持军事医疗人员的自动信息在当前的健康 在军事行动中遇到的士兵；

![](/media/202412//1733390409.2003958.png)它允许医疗后送小组在发生大规模伤亡时作出快速反应 战场；

![](/media/202412//1733390409.2101626.png)它支持住院前自动分诊；它支持更新电子战术战斗伤亡护理卡。.

该系统的局限性是，它需要一个工作的通信系统来定期地将测量数据从传感器传输到推理 和分析子系统。因此，它不能在所谓的无线电静音过程中工作。

本文的组织结构如下。下一节描述了关于常规医学和专门医学的类似健康监测系统的相关 工作。在第3节中，阐明了所提出的dss-医疗服务系统架构。该系统的要素从第4-7节中得到了 描述。最后一部分是论文的结尾。

2.相关工作

远程运行状况监控并不是一个新的技术问题。在过去的几十年里， [1]为可穿戴健康监测 传感器系统提出了许多解决方案。此外，许多现成的产品可以找到在市场上的[2 –5]。他们主 要由测量单元经常或定期使用的病人在家里测量和监测选定的生命体征，数据传输单元（主要 是智能手机或wi-fi路由器）和健康监测中心，允许可视化的威胁健康警报，测量生命体征和基 本生物医学信号。这些解决方案在所谓的远程医疗时代越来越重要，在这个时代，医生可以对 病人的健康问题迅速作出反应。如引言中提到的，远程健康监测主要由诊断患者利用。该解决 方案还配备了在识别出关键警报时自动呼叫紧急服务。不幸的是，在军事行动中，士兵不能直 接使用常规的健康监测系统来持续监测他们的关键生命体征。

目前市场上可用的远程健康监测解决方案是基于所谓的可穿戴设备，从军事的角度来看有 许多重要的限制，但重要的是，它们是专有的解决方案，因此不能轻易地与军事总部管理系统 集成。其中一个有趣的系统是和风生物利用公司的[6]，它也被制造商提出用于有限的军事用 途。不幸的是，它不允许测量我们假设在我们的系统中使用的所有生命体征(i。e., 血压，氧 饱和度)，但也有它的通信范围和技术是不可接受的战术行动（高达274米与和风回声网关通信 模块）。

在[7]中提出了一种有趣的远程监测足球运动员健康状况的方法，其中运动响应通过GPS ( 全球定位系统）单元进行记录，并将其传递到分析中，以根据速度强度优化运动员的表现。据 我们所知， 目前还没有研究使用类似的健康监测方法来评估作战效果或在军事行动中避免士兵 伤害的能力。我们也不认为我们的系统应该支持这种能力。避免士兵的伤害是非常重要的，但 这并不是军事行动战区的首要任务。因此，士兵的活动应主要进行优化，以达到军事任务。

[8]中提出的CRI（代偿储备指数）是一种新的生命体征参数，它使用了动脉脉冲波或光体 积描记术（PPG）的特征。在我们的系统中，我们也测量PPG信号，我们有能力远程将其发送到 分析和推理子系统。正如[8]的作者所指出的，CRI参数可以用于识别出血的创伤士兵，我们将 尝试在我们的系统的下一个版本中考虑这些参数。

反过来，在文章[9]中，作者探讨了三个参数（si-休克指数、pp-脉压和ROX指数），根据 它们在医院患者的病理生理学分类中的有用性。然而，这些都是医院的状况，而工作的结果是 关于死亡率的。我们的系统的主要作用是支持关于士兵从战场上撤离的决定。然而，上述参数 是根据主要生命体征i进行计算的。e., 心率，血压，氧饱和度和呼吸频率，这些也在我们的 系统中测量。

一些解决方案也专门用于支持军事医疗保健。其中一个解决方案是由美国开发的北约第一 反应器（NFR）应用程序。S.国防卫生局记录受伤和照顾伤亡在战场上补给[10]。该应用程序支 持准备报告

在受伤士兵被疏散到固定医疗设施之前，被送往医疗服务机构。在这种情况下，通信是基于移 动网络的。在缺乏访问蜂窝网络资源的情况下，关于患者的信息可以使用近场通信（NFC）标 签进行共享。上述解决方案符合北约对战场上士兵的疏散和医疗保护的标准。然而，值得注意 的是，NFR并不支持对受伤士兵的远程监测。

OpenAhlta是美国大学的一个开源版本。S.美国国防部战场电子病历系统（AHLTA-战区）

, 并使用NFR应用程序[11]进行操作。它使用健康级别7（HL7）标准来传输临床和管理数据。不 幸的是，它不能成为一个对战场上士兵远程健康监测的基础系统。

BATDOK（战场辅助创伤分布式观察工具包） 由美国创建和拥有。S.美国空军研究实验室的 第711人类性能翼是一款可以在智能手机或其他移动设备上运行的软件，使医疗人员能够在受 伤点[12]无线监测多名患者的生命体征。生命体征值可以通过所有可用的方法传递给应用程序 , 开始使用可访问的专门医疗设备进行测量，最后由急救人员对受伤士兵的观察。这个解决方 案在概念上与我们的系统非常相似，但它没有与传感器集成，甚至在受伤前就不断监测士兵的 健康状况。它也没有与战场管理系统集成，以不断支持不同指挥级别的医疗服务。

除了远程健康监测系统外，还阐述了一些分诊支持算法。其中大多数只是救援人员必须申 请的能够在事故现场有效提供医疗援助的程序。这种算法的一个例子是start（简单的分诊和快 速治疗） [13]或SALT（排序、评估、救生干预、治疗和/或运输） [14]，START算法支持第一反 应者，允许他们在很短的时间内，基于三个观察，i。e., 呼吸、灌注和精神状态。应答者将患 者分为四类：死亡的、即时的、延迟的和轻微的。分诊标签（主要是物理标记）被用来区分按 不同类别分类的患者。与开始类似，SALT是一个四步的过程，帮助急救人员管理大规模伤亡事 件。它还提出了这些标签来隔离患者（死亡、立即、预期、延迟和最小的）。START和SALT都是 基于不同的基本观察结果的相对简单的分类算法；因此，它们通常被用于分离过程的第一阶段 。更先进的系统，如TEWS（分类预警系统） [15]或NEWS2（国家预警评分2） [16]需要更多的测 量(i。e., 用血压、心率、血氧饱和度等)对伤亡人员进行分类。

[17]的作者，在他们的论文中，讨论了一种可以成功地应用于常规急诊服务的分诊方法。 该方法是指对随后需要住院且可能在24小时内死亡的患者进行分诊。 由于可用的标准指出引用 的论文可以用于选择士兵可以快速恢复和回到军事行动，我们提出几乎相同的方法在系统中所 谓的反向分类，但我们应该记住，战术操作经常阻止快速医疗援助和疏散，救援人员往往不能 立即到达受伤和疏散过程延迟（军事重点比军事行动中受伤士兵的医疗后送更重要）。我们拥 有不仅基于呼吸频率或氧饱和度测量的持续监测系统，甚至可以在到达受伤点之前，甚至在缺 乏医疗援助期间，为军事医务人员提供快速反应的手段。医疗后送人员可以根据来自我们系统 的信息，远程指导（使用军用无线电）第一反应人员（可以帮助受伤士兵的邻近士兵）。此外 , 我们监测的基本生命体征只建议进行最初的分诊，这必须由急救人员（能够准备和发送医疗 请求的指挥官或士兵）来确认

疏散)，在引用论文中指出的容易获得的标准。除了关于初始分诊的信息外，我们的系统还允许 发送某些生物医学信号，这些信号是根据初始分诊所提醒的医务人员的要求而产生的。 由于对 选定的生命体征的持续监测，上述所有行动都可以自动启动。然而，应该注意的是，我们的系 统也有一些局限性。这种情况将发生在无线电静音期间（由于安全原因，无线电不能传输数据 的特定军事情况），但这种限制适用于所有远程监控系统。

考虑到上述的分诊支持系统，作者准备了一套dss-医疗治疗系统必须实现的能力。第3节 描述了医疗系统所需的能力。

3.为的决策支持系统的建议架构和要求

医疗疏散（DSS-MEDEVAC）

在军事行动中支持医疗人员的系统将不同于典型的医疗远程监控系统，尽管这种系统在这 两种情况下的目标是相似的；他们必须测量病人/士兵/消防员的相关生命体征，并将数据转移 到健康监测中心，以便更快地对关键的健康事件作出反应。然而，在军事、战术环境下（在战 场上），首要任务是实现军事目标。因此，快速恢复士兵以便在战斗中再次使用它们是医务人 员的一项关键任务。为了应对这一任务，在院前医疗护理中经常使用反向分诊。

为了帮助军事医务人员决定快速撤出战场，需要持续的健康监测，而不是关注士兵的特定 疾病（但考虑到个人健康参数）。这种持续的监测不仅需要对人类生命体征进行可靠的测量，

还需要考虑到特定的战场条件（士兵可以奔跑、下降、可能压力过重，最终可能轻微或重伤， 或可能在战场上死亡） 。考虑到这些条件，士兵的健康监测系统将不同于民用市场上可用的健 康监测解决方案。作者准备了这种系统的架构，重点是在战场上支持医务人员，特别是在医疗

后送过程之前和过程中更快的分诊方面。在系统描述中使用的缩写收集在 表1。

表1。在dss-医疗服务系统描述中使用的缩写的列表。

缩写

阿姆布

羊水栓塞

AIC AIS 轮廓 禁止

骨髓单核细胞 BMS

意义

空中救护

模拟前端分析和推理能力分 析和推理子系统

救护车

身体区域生物传感器网络 战场医疗监测中心

战场管理系统

车身位置

情报和监视

BodyPos 命令、控制、通信、

c3是

指挥官节点

cn

中央处理器 DMN

DTC

数据传输服务 出口信贷保证

中央处理单元

决策节点

数据传输能力

数据传输子系统

心电图

表1。续。

缩写

意义

gRPC HMS

人力资源

冰点 MDTP

梅格 人

MMC

MPA

马萨克

MTF NAFv4

非视力损伤情况

PAS PAT 波伊 PPG

邮电总局 雷斯

rr SBC sn

SpO2, SpO2 SVM

塔台管制计算机组 阿尔特

统一的S波段 SBP

长春新碱

谷歌远程过程调用

总部管理系统

心率

互联网协议 医疗数据通信数据传输协议

医疗疏散小组 医疗疏散节点 医疗监测中心

平均体力活动

医疗状况意识能力医疗设施 北约架构框架版本4

嵌套的矢量中断控制器 体育活动信号

脉冲到达时间 受伤点

光电体积描记图 脉冲传输时间

呼吸频率信号 呼吸频率

单板计算机

士兵节点

氧饱和度 支持向量机

战术战斗伤亡护理通用异步接收机-发射 机

通用串行总线 收缩压

虚拟COM端口

3.1.系统要求

专用于军事行动的dss-医疗服务系统的体系结构是基于北约体系结构框架版本4（NAFv4） [18]准备的，这是一个用于开发和描述军事和商业用途的体系结构的标准。NAFv4支持所开发系 统的不同视点的准备，包括系统概念（C组视点）、服务规范（S组视点）、逻辑规范（L组视点 ) 、物理资源规范（P组视点）和架构基础（一组视点）。

在本节中，描述了dss-医疗服务系统的主要观点。e., 能力分类（C1）、企业视觉（C2） 、能力依赖性（C3）、服务分类（S1）、节点类型（L1）和逻辑方案（L2）。

为了反映医疗系统的目标，定义了主要系统能力。图1显示了提出的简化的系统能力分类 （C1）和能力依赖（C3）观点。

据假定，在确定（检测到的）生命或健康威胁（医疗状况意识能力-MSAC）的情况下，该 系统必须向医务人员和选定的指挥官提供关于士兵生命体征的当前信息和关于分诊的初步决定 。此外，该系统必须提供每个士兵的当前位置，以帮助组织医疗后送。该系统必须能够测量、 登记和处理所需的人体生命体征（生命体征监测能力），并可靠地将测量结果的结果（数据传 输能力-DTC）发送到分析和推断模块。然后，系统必须有效地处理测量数据，以提出分诊的初 始决策（分析和推理能力-aic）。

![](/media/202412//1733390409.2299209.png)

图1。美国医疗服务中心的能力分类和能力依赖关系。

经过详细分析，确定系统的主要目标可以通过以下人体生命体征(即：心率（HR）、呼吸 频率（RR）、氧饱和度(SpO2)、平均体力活动（MPA）、体位（BodyPos）和收缩压（SBP）。

生命体征是根据四个因素选择的。e. :

.支持批量诊的解决方案的相关工作分析

人员伤亡事故（例如： [19,20])；

![](/media/202412//1733390409.2395701.png)关于在灾害期间支持远程健康监测的相关工作分析- 通过Thig s互联网（IoT）解决方案诱发的大规模伤亡事件（例如： [21]）；

![](/media/202412//1733390409.2494051.png)分析军方专用的类似解决方案（例如： [10-12]）；急救人员和院前医务人员的个人医疗经 验

在军事行动期间进行撤离。

[21]的作者提出了一种实时监测电子分诊标签系统，用于提高灾难引起的大规模伤亡事件 中的存活率——因此，该系统的目标与我们的系统的目标相似。他们还提出了在院前分诊时可 以简单测量的生命体征。e., 体温、心率、血压、呼吸频率、氧饱和度、毛细血管再充血、意 识、心电图。 [21]的作者还认为，心率、血压、呼吸频率和心电图等生命体征被现场救援人员 认为是最重要的因素。 由于我们系统中的分诊是基于简单的分诊和快速治疗（开始）程序，我 们还提出了一套可以持续、 自主、无创、重要的是、简单可测量的生命体征

在军事行动期间(例如，SpO2即使没有专门的医疗设备（例如，RR）。

我们系统的生命体征的选择也来自于对其他分诊系统（程序）的分析。大多数分诊系统是 基于简单的意识测试，使用ACVPU（警报、混淆、声音、疼痛、无反应）量表、桡动脉脉搏和 呼吸次数，但在更先进的分诊系统中，SpO2、SBP、HR。在[19]中，显示了对目前正常运行的 分诊系统的广泛分析。 [20]的作者还对与大规模伤亡生物恐怖主义有关的分诊系统和有用的生 命体征进行了分析。我们为系统选择的生命体征通常是分析分诊系统中考虑的主要变量。该分 析使我们得出结论，我们需要一套最小的生命体征，能够考虑到强有力的军事限制和程序，支 持对士兵健康状况的快速远程推断。

AIC还应由每个士兵专用的校准和参考数据提供（校准和个人参考数据加载能力） 。此外 , 该系统必须使医务人员或救援人员能够远程观察选定士兵所需的生物医学信号，以帮助他们 提前准备所需的医疗设备和药物，并在需要时更新分诊决定（生物医学信号监测能力）。根据 远程监测士兵的生物医学信号：心电图（ECG）、光体积描记图（PPG）、身体活动信号（PAS） 和呼吸频率信号（RES）。受伤士兵登记的生命体征必须自动传递到电子战术战斗伤亡护理卡， 该卡必须立即在系统中供参与士兵治疗的医疗人员使用（电子TCCC卡填充能力）。TCCC卡和最 初的分类决定都必须立即在军队在军事行动中使用的总部管理系统（HMS）中“实时 ”查看（可 视化能力）。

3.2.系统需求的实现

服务分类法提供了一个观点（S1），如图2所示。DSS-MEDEVAC系统提供的服务反映了所需 的能力（如图1所示）。主要的服务组负责可视化与分诊决定、士兵的定位、生命体征、生物 医学信号和电子TCCC卡有关的信息。系统所需要的重要服务也是数据传输服务，它确保了系统 中所有参与者的可靠的数据分布。

![](/media/202412//1733390409.2620542.jpeg)图2。DSS-MEDEVAC服务分类法的观点。

图3显示了计划使用dss-医疗服务系统的医疗后送行动的操作观点。受伤的士兵首先被疏 散到战场上的安全区域（如果可能的话）给他们提供急救。让我们将这个区域命名为受伤点，

在那里医疗人员（军事救护车-amb或空中救护车-AAMB）可以提供第一个专业医疗援助。DSS-医 疗系统的传感器应首先对伤亡人员的生命问题作出反应，医务人员在疏散初期得到重要生命参 数的支持。还建议立即进行初步分诊；因此，所需的医疗资源可以参与医疗后送的过程。此外 , 还可以在撤离前和在途中护理（支持专业医疗器械）期间观察到受伤士兵当前的生物医学信 号。图3中指出的军事疏散的上层可以利用电子TCCC卡和初始和更新的分类信息以及受伤士兵的 当前情况。操作观点显示了dss-医疗治疗系统的主要参与者。

![](/media/202412//1733390409.2707508.png)

图3。系统的运营（企业）观点。

图4阐明了dss-医疗服务系统的一般观点。在单个士兵级别，必须使用一组生物传感器来 测量生命体征和生物医学信号，这对于最初的分诊准备和所有医疗后送过程的角度来看（从能 力的角度来指出和解释）很重要。

战场管理系统（BMS）将士兵生命状况的一般信息（初始和最终分类）可视化给直接指挥 官，支持当前的态势感知。使用准备好的通信协议，通过通信网络不断发送到医疗监测中心（M MC），在那里对数据进行分析，提出分诊的初步和最终决定。然后，将这些信息传递给MCC的医 务人员可用的可视化模块，该模块可以对医疗后送作出最终决定。医务人员还可以观察受伤士 兵所需的生物医学信号，以进一步支持有关撤离的决定，或就急救问题提供远程建议。同样的 信息也出现在医疗后送小组（MEG）-救护车的可视化模块中。MEG的医疗人员可以在受伤点采取 直接医疗行动后改变分诊决定。最初填写了电子表格

与每个受伤士兵相关的TCCC卡也由救援人员更新，并立即向疏散设施各级的医务人员提供。

![](/media/202412//1733390409.282373.png)

图4。分布式医疗服务系统的一般观点。

dss-医疗治疗系统的逻辑观点如图5所示。系统中有三个主要节点：士兵节点（SN）、指 挥员节点（CN）、决策节点（DMN）和医疗疏散节点（MEN）。

SN由身体区域生物传感器网络（BAN）、单板计算机（SBC）、个人终端和个人无线电组成 。BAN是一个与士兵的个人装备（包括军装、头盔和内衣）集成的网络。SBC旨在执行生物医学 信号的初始处理和生命体征的实时计算，并通过已开发的通信协议将处理后的数据提交给个人 终端。个人终端是一种基于平板电脑的军事计算机，它构成了通往战术数据传输系统的门户。

人们还认为，BAN，特别是SBC，可以监测和存储士兵的生命体征的历史，即使其他设备无法接 近。如果需要，可以稍后访问这些值。战术数据传输系统是BMS和HMS系统的一部分。它的主要 作用是使用战术无线电网络（个人无线电、车辆无线电设备、部署和固定通信基础设施）可靠 地分发医疗通信系统的所有元素上的医疗数据。

![](/media/202412//1733390409.2952785.jpeg)图5。地中海逻辑视点-节点类型视点。

CN的角色是收集由下属传输的医疗数据，并将其发送给DMN。基于从CNs接收到的数据，

DMN节点执行高级分析和推断，并提出关于分诊的决策。该决策被传递给位于指挥官节点、医 疗后送节点和参与医疗后送行动的其他医疗设施中的可视化模块。医疗数据通过战术数据传输 系统分布在dss-医疗网络的节点之间。关于分类、生命体征和生物医学信号的信息是

使用BMS和HMS系统向医务人员和指挥官进行可视化。医疗后送节点（救护车、航空医疗直升机 ) 接收当前的分诊决定，并首先填写电子TCCC卡，以支持救援人员。男性也是最终决策信息的 来源，这些信息被传递给DMN并在系统中分发。

DSS-MEDEVAC系统的主要子系统，它们反映了图5中所示的逻辑视点的组成部分，将在下一 节中进行描述。

4.测量子系统

使用不同类型的传感器来测量在相对静态的身体位置的生物医学信号和生命体征在文献中 有很好的描述。在DSSMEDEVAC系统中，我们假设了困难的测量条件，其中一些传感器可以被破 坏或断开，对接收数据的分析应该可以预测这种情况。此外，我们应该假设，通常已知的无线 传输技术专用于传感器网络(i。e., 蓝牙、ZigBee、ANT或近场通信)不能在军事行动中应用，

特别是由于故意干扰无线电传输。所有这些考虑都导致我们决定了一组应该被监控、最初处理 并发送到分析和推理子系统的健康参数，以及决定了防止数据传输被干扰的通信技术。此外， 可穿戴传感器应与个人士兵的装备（制服、 内衣、通信设备）集成。测量子系统的简化体系结 构如图6所示。

![](/media/202412//1733390409.3075213.jpeg)图6。医疗系统测量子系统的总体结构。

由本文作者制造的可穿戴传感器位于一名士兵的前额、胸部和手腕（可选的传感器）上。

它们与头盔和制服相结合。基于高性能手臂皮层的m7单板计算机（SBC）负责传感器的数据采集 、测量信号的初始处理、生命体征的实时计算和与个人终端的通信。为了确保干扰保护，数据 通过有线通信传输到SBC，其中电线要统一集成。

SBC和传感器都由个人终端（作为主要能源）或内部SBC电池（处于备用或紧急模式）供电。假 设个人终端在军事任务期间能够向测量子系统传递所需的能量。数据通过个人无线电发送到系 统的其他部分，以确保安全和健壮的战术通信（即使是在敌对干扰下）。

一般来说，可以测量大量的生物医学信号和生命体征，以支持远程健康监测[2,22]。然而 , 我们不需要使用所有这些方法来支持关于在军事行动（或类似行动）中的分类的决定。考虑 到dss-医疗治疗系统所需的能力，特定的军事条件，以及在对军事行动期间的军事医疗保健和 疏散程序[23]进行深入分析后，我们决定测量图1中指出的生物医学信号和生命体征。这些标志 是由关于士兵的运动和身体位置的信息（通过使用一个加速度计）来支持的，如图6所示。在系 统开发的当前阶段，作者准备了一个包含所有所需传感器的测试台，其中每个传感器由两个主 要组件组成：CPU（中央处理单元）和AFE（模拟前端）模块，如图7所示。传感器能够通过双线 UART（通用异步接收器-发射器）接口与SBC进行通信。这种传感器的模块化架构允许为具有相 同CPU模块的每个传感器使用专用的AFE。

![](/media/202412//1733390409.3223884.png)

图7。dss-医疗服务传感器的架构。

如图6所示，DSS-MEDEVAC的测量子系统包含三个测量生物医学信号的传感器。这些信号集 成在SBC计算机中；因此，确保这些信号的测量同步是很重要的。这在计算SBP参数时尤为重要 。SBP将根据从ECG/PPG信号[24–27]计算出的脉冲波的传播时间（PAT-脉冲到达时间或PTT-脉 冲通过时间）进行连续、无袖和无创的测量。在我们的测量子系统中，如图6所示，我们提出了 一个由SBC同步的分布式有线传感器系统。为了控制传感器，SBC会定期向每个传感器发送请求 。每个传感器都必须响应该请求(i。e., 它在一个固定的时间内发送信号采样)，小于请求信号 的最高采样周期（例如，对于ECG，响应时间必须小于4 ms）。在我们的传感器中，我们使用了 STM32L5系列的微控制器，它们嵌套了矢量中断控制器（NVIC），允许准备多达8个中断优先级 。传感器从SBC接收到的请求(i。e., 从UART接口接收一个字节的中断)具有适当的高优先级。

由于这个原因，传感器的响应(i。e., 通过UART开始数据传输)的时间低于100秒。μ图8显示， 所开发的AFE模块能够测量和发送心电和PPG信号，然后将其发送到SBC。

![](/media/202412//1733390409.3345757.jpeg)图8。 由传感器的AFE模块测量的示例信号图（ “ ， ”表示十进制分隔符）。

使用信号植物[28]软件来显示图8中的数据。它便于以相同的采样率查看多个信号，并具 有许多有用的功能，包括数据处理。

5.数据传输子系统

数据传输子系统（DTS）的一个作用是为DSS-MEDEVAC系统提供数据传输服务（见图2）。 DTS集成了SN、CN、DMN和MED节点，如图9所示。

![](/media/202412//1733390409.3468513.jpeg)图9。数据传输子系统作为dss-医疗服务节点的集成器。

总部管理系统（HMS）确保了节点之间可靠的医疗数据分布，该系统最初开发用于支持军 事行动[29]期间的指挥、控制、通信、情报和监视（C3IS）能力。HMS C3IS系统的基本形式使 指挥官能够及时地做出决策，通过获取报告的信息，创建作战情况的共同图景，并交换作战信 息。因此，HMS还可以通过dss-医疗中心的信息，以支持军事医疗保健各级的医疗人员，包括军 事医疗后送活动（如图4所示）。DTS的一个核心是一个基于互联网的战术通信网络

协议（IP）。在基于ip的通信网络的顶部工作的HMS确保了可靠的数据库更新和复制。在本节 中，我们重点关注DTS的一部分，i。e., 关于数据传递到HMS的方法。

图10显示了士兵节点的一个通信组件。

![](/media/202412//1733390409.3656611.png)

图10。士兵节点的通信组件。

如前所述，传感器由SBC通过UART接口进行控制。接下来，数据使用通用串行总线（USB） 接口传输到个人终端，其中SBC以设备模式运行USB驱动程序，而个人终端以主机模式使用USB驱 动程序。为了协调通过USB接口的数据传输，在SBC（MDTP设备模式）和终端（MDTP主机模式）

中，制定并实现了一个MDTP协议（DSS-地中海数据传输协议） 。MDTP驱动程序使用虚拟COM端口 （VCP）与USB驱动程序进行通信。数据通过谷歌远程过程调用（ gRPC） [30]从HMS C3IS系统发 送，该[30]通过IP数据包在gRPC客户端和服务器应用程序之间传输。gRPC客户端是集成器的一 部分，它将数据分布在dss-医疗服务节点上。Iv接口负责进程间的通信。

6.分析和推理子系统

在本节中，只提供了关于分类算法的一般信息，这可以阐明分析和推理子系统（AIS）在 DSS-MEDEVAC系统中的作用。AIS的一个核心是一种决策算法，通过将受伤的士兵分配给他们以 下颜色：绿色、黄色、红色和黑色。前三种颜色反映了在开始（简单的分诊和快速治疗）程序 中定义的分诊要求。最后一个是为了表示通信问题。

我们还根据对医疗后送人员的经验所造成的实际情况的评估，提出了士兵生存机会度量的 适当值。这些价值将在武装部队实施该系统后的经验教训进行优化。 目前，我们在最初的现场 测试中验证了我们的假设，证实了系统的所有元素都可以与战场管理系统集成，但由于明显的 原因，我们无法在有大量受伤士兵的真实战场上进行验证。因此，我们假设系统能够学习和修 改其初始属性。

绿色意味着一个士兵的生存几率等于100%。这意味着士兵确实可能受轻伤，但不需要医疗 援助。黄色被分配给应该撤离的受伤士兵，但第二优先。红色保留给需要立即撤离的士兵（优 先考虑） 。如果没有从士兵的测量设备接收到心率信号（HR），则会分配黑色。这可能意味着 通信问题累积或士兵死亡（这种颜色需要医务人员或其他士兵的额外验证）。除了上面定义的 颜色外，蓝色也可以分配给没有（或最小）生存机会的重伤士兵。该颜色只能由医疗人员或来 自医疗后送组的救援人员手动分配，

在确认了士兵在受伤时（通常是在战场上）的健康状况之后。根据军事程序，这些士兵将以最 后的优先顺序撤离（称为撤退分流）。

第3.1节中描述的系统要求要求AIS应由四个主要生命体征提供，以作分诊决定，i。e., 心率（HR）、呼吸频率（RR）、收缩压（SBP）和外周血氧饱和度(SpO2).该算法还可以由关于 身体活动和身体位置的信息（由加速度计估计）来支持。在医疗隔离的背景下，士兵健康评估 的最高优先事项是人力资源。RR是一个对人类健康状况的变化产生快速反应的参数。SBP是HR 和RR的重要体征。该参数的值在患者病情恶化时反应较慢。SpO2不如以前的参数可信，因为它 取决于许多因素，包括那些与生命和健康威胁没有直接联系的因素；因此，在最后准备分诊决 定时应考虑到这些因素。

AIS不需要由测量子系统测量的生物医学信号。它们按需直接送至位于医疗监测中心的医 务人员或接近事故现场或战场的救援人员。因此，生物医学信号支持急救和对分诊的最终决定

。

主要的分类算法用算法1中所示的伪代码来澄清。根据它，如果所有被测量的参数都在个 性化参考值的范围内，那么（绿色）颜色将被分配给士兵。如果至少有一个参数的值超出参考 范围，但不超过临界值（表示对运行状况的威胁），则将指定（黄色）颜色。该算法的规则之 一是，如果更重要的参数的值强制使用黄色或红色，则分类颜色不能改变下来(i。e., 分别为 绿色或黄色)。分析后，最重要的生命体征是RR，然后是HR、SBP，最后是SpO2.因此，在这样 的顺序下，在算法1中分析了符号(第1行的RR，第8行的HR，第23行的SBP和SpO2从第38行）。

此外，主要算法还补充了一套算法，准备评估受伤士兵的生存机会，并对传感器缺乏一些 参数（在军事行动中经常观察到）作出反应。

在主算法的基础上，定义了一个确定生存机会功能的辅助算法。定义这一个功能的需要来 自于与引入支持分诊的软件相关的期望。疏散的顺序来自于指定的状态，用适当的颜色标记，

但在相同的颜色内，生存机会功能的价值可能是决定性的。

在所谓的战术反向[31])分类的情况下——作为TCCC（战术战斗伤亡护理）的一部分—— 根据战术情况，当优先事项是尽快恢复执行战术行动的能力时，使用它，因此最轻的伤员将首 先被营救；因此，他们将能够尽快返回战术行动， 以重建作战能力。这是一种特殊的情况，即 拯救受伤最少的人，以便能够在战斗中再次使用它们。在这种情况下，生存机会功能的价值也 将非常有帮助。一般来说，对于救援者来说，关于输入的信息越多，生存机会的功能就越好— —口语中说——可以让你评估哪一种红色可能比其他的“更 ”红色，这可以改善医务人员在优 先疏散方面的行动。

为了定义生存机会函数，我们生成了7200个病例，它们很好地覆盖了整个四维参数空间。 这些案例被用于训练两个非线性SVM（支持向量机）网络[32 –34]，最终的结果是一个函数， 分配红色类的范围为1-50%，黄色类51-99%，绿色类100%。

算法1是在分析和推理子系统中实现的主要分类算法。

输入：

RR、HR、SBP、SpO2的测量值。

对每个士兵的生命体征的个性化参考和临界价值。 典型参考值：

RRref: 9 –20/min

HR ref : 50 – 110/min SBPref： 100-180mmHg SpO2ref: >= 94%

RRctitical, HRcritical, SBPcritical, SpO2Critical 输出：

治疗类选法

步骤：

1.读取RR

2.如果RR == RRref，那么 3. |分类<-（绿色）

4.否则如果RR ！=公司 5. |分诊<-（黄色）

6.其他的

7. 分诊<-（红色）

8.阅读人力资源

9.如果分诊==（绿色）

10. |，如果HR == HRref，那么 11. ||分类<-（绿色）

12. |其他如果人力资源！=关键 13. ||分类<-（黄色）

14. |其他

15. |分诊<-（红色）

16.否则，如果分类==（黄色） ，然后 17. |如果HR == HRref或HR！=HR关键字 18. ||分类<-（黄色）

19. |其他

20. |分类<-（红色） 21.其他的

22. 分诊<-（红色） 23.读取SBP

24.如果分诊==（绿色）

25. |，如果是SBP==，SBPref，那么 26. ||分类<-（绿色）

27. |其他如果SBP！=SBPcrect 28. ||分类<-（黄色）

29. |其他

30. |分类<-（红色）

31.否则，如果分类==（黄色） ，然后

32. |，如果SBP == SBPref或SBP！=SBP关键字符，然后

33. ||分类<-（黄色） 34. |其他

35. |分类<-（红色） 36.其他的

37. 分诊<-（红色） 38.读取SpO2

39.如果分诊==（绿色）

40. |，如果SpO2，==SpO2ref，那么 41. ||分类<-（绿色）

42. |其他如果SpO2 ！=SpO2关键

43. ||分类<-（黄色） 44. |其他

45. |分诊<-（红色）

46.否则，如果分类==（黄色） ，然后

47. |如果SpO2 == SpO2ref或SpO2！=SpO2关键

48. ||分类<-（黄色） 49. |其他

50. |分类<-（红色） 51.其他的

52.分诊<-（红色） 53.算法的结束

为了训练SVM网络，我们使用了由受伤患者生命体征的高保真模拟器生成的数据。该模拟 器通过测量创伤患者的健康参数进行了验证。这一决定是由于缺乏对真实患者进行的足够数量 的测量，并且仅用于验证我们的算法。但是，需要强调的是，SVM网络将在系统的最终运行过程 中进行训练，以优化其行为。红色、黄色和绿色等级范围应视为初始值，计划根据军事经验进 行优化。

所选参数的生存函数机会图如图11所示。应当指出的是，所提出的功能也将根据所吸取的 军事经验教训进行优化。我们可以预期，在我们的系统运行的初始阶段，可能会出现错误的警 报，但使用学习能力，它应该调整其性能。

![](/media/202412//1733390409.4105191.jpeg)图11。选定的生命体征值的生存机会函数图。

在战场上，一个或多个传感器被切断并不罕见。这可能是由于信号截止或强干扰阻碍了参 数的正确确定。因此，使用了以下系统操作的变体：

1 . 所有四个参数都可用的。使用了完整的算法，仅在这种情况下，才计算了生存机会函数的值 。可靠性为100%。

2.如果SpO2传感器断开连接，仅采用HR、RR和SBP的算法。可靠性为90%。

3.如果SBP传感器断开，则只使用HR、RR和SpO的算法2是应用。可靠性为80%。 4.如果SpO2并断开SBP传感器，仅采用HR和RR的算法。可靠性为70%。

5.如果RR传感器断开，则仅使用HR、SBP和SpO的算法2是应用。可靠性为80%。 6.如果RR和SpO2传感器断开，使用只使用HR和SBP的算法。再负债为70%。

7.如果RR和SBP传感器断开，则该算法只使用HR和SpO2使用。可靠性为60%。 8.如果RR，SpO2 ，并断开SBP传感器，使用仅hr算法。可靠性为50%。

9.在没有来自HR传感器的信号的情况下，系统不能工作（不管其他的）。士兵被标记为“黑色

”, 但有必要由现场的医生确认。 AIS的总体架构如图12所示。

![](/media/202412//1733390409.4252477.png)

图12。分析和推断子系统架构。

AIS算法在分析和推理服务器中实现，该服务器通过谷歌远程过程调用通过IP接口与HMS C3IS系统进行通信。AIS算法由每个士兵的生命体征提供。为了将当前的计算与历史数据进行 协调，参数值和结果都存储在一个数据库（DB）中。每次改变分类颜色时，关于分类的信息都 会被发送到HMS C3IS系统。HMS C3IS系统负责向指挥官（以有限的形式，需要指挥和控制）、 医务人员或救援人员可视化分类。

图13显示了一个示例窗口，其中包含了对选定士兵的分类可视化的建议决策。

![](/media/202412//1733390409.4742124.jpeg)图13。 由AIS算法生成的建议的分类颜色的示例窗口。

根据传递给AIS的参数，算法建议士兵应该用（黄色）颜色进行标记。所有的生命体征都 可以到达AIS处；因此，结果的可靠性是100%的。SVM网络显示，该士兵的存活几率为74%。关 于医疗后送和分诊的最后决定现在由医务人员作出，他们还可以另外检查士兵所需的生物医学 信号。

7.针对医务人员和指挥官的可视化模块

为了使医疗支持小组（MSG）/战场医疗监测中心（BMMC）的运营商收集的数据可视化，设 计和实施了专用门户，作为波兰总部管理系统（HMS C3IS茉莉） [29]的一部分。该门户提供了 对由系统监控的有关士兵的信息的访问。适应于T5“TRYTON战术计算机终端的MSG门户视图如 图14所示。主要视图提供了允许医疗人员或指挥官了解战场情况的基本数据：士兵名单、分类 的最新结果和关于数据更新的信息。

![](/media/202412//1733390409.4868205.jpeg)图14。味精门户：视图适应于T5“TRYTON战术计算机终端。

使用该软件，味精操作员可以访问每个被监控士兵的详细数据。 门户提供获取以下数据：

![](/media/202412//1733390409.5236664.png)基本信息：性别、出生日期、体重、最新的心率和血压 消息

![](/media/202412//1733390409.5343583.png)生命体征的状态，i。e.: 呼吸频率，氧饱和度，温度和信息- 关于身体活动和体位的问题；

![](/media/202412//1733390409.5434873.png)一个动态更新的图表，反映了选定的生命体征随时间的变化；

![](/media/202412//1733390409.5531294.png)分类结果：从分析和推理模块（初始分类）和 由操作员或医务人员手动输入的信息（最终分类）。

“分类 ”选项卡显示了从分析和推理服务中读取的结果列表。医疗支持小组门户还允许使 用地图层可视化数据，为救援人员提供更大的态势感知，从而促进未来的受害者疏散计划。作 为基本信息的符号学表示每个被监测士兵的分类结果。此外，操作员还可以显示士兵最后登记 的生命体征的摘要。

“信号 ”选项卡允许可视化由选定士兵的测量模块注册的选定生物医学信号。图15显示了 其中一个测试生物医学信号(i。e., photoplethysmogram).

MSG门户网站的设计使用了响应式视图技术，允许界面适应屏幕的分辨率和方向。这也允

许与移动设备的高效工作，e。g., TRYTON的“T5战术计算机终端 ”，它被适应于 在任何地形上工作，在各种环境条件下工作。

![](/media/202412//1733390409.563846.jpeg)**图15。**MSG门户：一个生物医学信号的可视化（士兵名称： Kowalski，信号名称：光体积图nr 1）。

**8.结论**

目前，DSS-MEDEVAC系统是以集成测试台的形式编写的，其中所有的元素都被实现和测试 。测试证实，准备好的传感器以分析和推断子系统和医务人员所要求的可接受的形式测量了生 命体征和生物医学信号。将生物医学信号与专业医疗设备获得的数据进行了比较。救援人员评 估，它们对受伤士兵的远程健康监测非常有用，是dss-医疗治疗系统建议的分类信息的一个很 好的补充。接下来将在现场测试中评估系统的效率，以确认所有所需的能力。

作者贡献：概念化出版社，P。L., J.K., T.S.和A。P.D. ; 方法，P。L., J.K., A.P.D.和P。O.；软件 , 。P.D., L.A.,J.K., T.S.和W。Z.；验证，P。L., J.K., T.S., A.P.D.,P.O., L.A.和W。Z.；正式分 析，P。L., J.K., T.S.和A。P.D. ; 写作-原稿准备，P。L., J.K., T.S., W.Z.和A。P.D. ; 写作-审 查和编辑，P。L., J.K., T.S.和A。P.D. ; 监督，J。K.；项目管理、P。L.所有作者均已阅读并同意了 该手稿的出版版本。

资助：该研究由波兰国家研究和发展中心（NCBR）资助，项目为No.DOB-SZAFIR/09/B/006/01/2021.

**机构审查委员会的声明：不适用。**

知情同意声明：不适用。

**数据可用性声明：不适用。**

**利益冲突：作者声明没有利益冲突。**资助者没有参与研究的设计；在数据的收集、分析或解释中；在手稿 的写作中，或决定发表结果。

**参考文献**

1.Majumder，S.；蒙达尔，T.；迪恩，M。J.用于远程健康监测的可穿戴传感器。传感器2017、17、130。CrossRef

2.《医疗保健行业中远程患者监测的技术、设备和好处》，《2021年远程患者监测趋势和健康设备》。在线可用： https://www。

insiderintelligence.com/insights/remote-patientmonitoring-industry-explained/（已于2023年4月24日访问）。

3. 维德曼，D.；塞里；E.；Oubakirova，M.；Abdildin，Y.； 巴登斯，R.；Bilotta，F。慢性危重症患者出院后的远程监测：系统回顾。 J.Clin。医学2022, 11, 1010.CrossRef

4.塞沙德里，D。R.；李，R。T. ; Voos, J.E. ; Rowbottom, J.R. ; Alfes, C.M. ; Zorman, C.A. ; 德拉蒙德住所名称 C.K.用于监测运动员 内部和外部工作量的可穿戴传感器。NPJ数字。医学2019, 2, 71.PubMed

5.查特吉，T.； 巴塔查里亚州， D. ; 人名 A. ; 朋友，M。远程军事训练活动的快和慢完成者的生理和心理工作负荷的量化。BMJ Mil.健康 2022 ，e002154 。PubMed

6.金，J。H.；罗伯奇，R.；鲍威尔，J。B. ; 叶形， A.B. ; 威廉姆斯，W。J.使用和风生物组织在分级运动和持续高温运动中测量心率和呼 吸频率的准确性。Int。J.运动医学。2013, 34, 497 –501.PubMed

7.罗素，M.；Sparkes，W.；东北，J.；库克，C.J。；爱，T.D.。；布拉肯，R.M.。；基尔达夫，有限公司。在整个职业足球比赛过程中， 加速和减速能力的变化。J.强度金度。物品2016, 30, 2839 –2844.PubMed

8.约翰逊，M。C. ; Alarhayem, A. ; Convertino，V.；卡特，R。，第三；钟，K.；斯图尔特，R.；迈尔斯，J.；凹陷， D. ; 廖；Cestero；

R；等。补偿性储备指数：一种新型监测技术识别出血创伤患者的性能。震惊2018,49,295-300。CrossRef

9. 凯莱特，J。；荷兰，M.；Candel，B。G.J.利用生命体征将急性病患者快速和容易地纳入临床有用的病理生理学类别：在两个不同的急性 病患者群体中的8个病理生理学类别的衍生和验证。J.艾默格。医学2023, 64, 136 – 144.CrossRef

10 . 北约急救反应APK，国防卫生局。可在线获得： https://apkcombo。com/natofirst-responder/mil.dha.第一个响应者（已于2023年5月 15日访问）。

1 1 . OpenAhlta。在线可用： https://www。健康mil/Military-Health-Topics/Technology/SupportAreas/Open-Source-Technologies（已 于2023年5月15日访问）。

12.BATDOK，空军研究实验室。可在线获得：https://技术lick中心。org/technologies/battlefield-assisted-traumadistributed- observation-kit-batdok-software-tools/d6074edb-dde4-4a0b-8c40-c4e844917b28（已于2023年5月15日访问）。

1 3 . 开始成人分诊算法。辐射应急医疗管理：REMM（美国卫生与公众服务部） 。在线可用： http://www。remm.nlm.gov/startadult.htm (

已于2023年4月24日访问）。

14.联邦紧急医疗服务跨间委员会。全国实施大规模伤亡事故分类模型统一核心标准。2 0 1 4 . 在线可用： https://www。

ems.gov/assets/National\_Implementation\_Model\_Uniform\_Core\_Criteria\_Mass\_ Casualty\_Incident\_Triage\_Mar2014 .pdf（已于 2023年4月24日访问）。

15.Hinson, J.S.；马丁内斯，D。A. ; 卡布拉尔；乔治，K；惠伦m；Hansoti B. ; 莱文，S。在急诊医学中的分诊表现：一个系统的回顾。安 艾默格。医学2019, 74, 140 – 152.CrossRef

16.肯普，K.；Alakare，J.；哈约拉，V。P.；斯特兰德堡，T.；托洛宁，J.；列L.；Castren，M。国家早期预警评分2（NEWS2）和3级分诊 量表作为急诊科中虚弱老年人的风险预测指标。BMC Emerg.医学2020, 20, 83.CrossRef

17.Kikomeko, B. ; Mutiibwa, G. ; Nabatanzi, P. ; Lumala, A. ; 凯莱特，J。A.一种在低资源环境下使用的急诊患者分诊工具的前瞻性、 内部验证。Afr。J.艾默格。医学2022, 12, 287 –292.CrossRef

18.架构能力团队，咨询，指挥和控制委员会，北约架构框架版本4,1月

20 18 . 在线可用： https://www。北约。in t/n ato\_ static\_ fl20 14/assets/pdf/202 1/ 1/pdf/NAF v4\_2020 .09 . pdf(访

问时间

2023年4月24日)。

19.Bazyar，J.；Farrokhi，M.；汗基，H。大规模伤亡事件和灾难中的分诊系统：一种具有世界性方法的综述研究。开放的访问麦克。J.医 学科学。2019, 7, 482 –494.CrossRef

2 0 . 赵F.；赵C.； 白S.姚L.张Y。大规模伤亡生物恐怖主义的分诊算法：一个系统的综述。Int。J.环境。物品公共卫生部门，2023,20,5070 。PubMed

2 1 . 公园，J。Y.实时监测电子分类标签系统提高灾害性大伤亡事故的生存率。医疗保健2021,9877。PubMed

22.波利的变体 C. ; 贾亚拉特纳；汉密尔顿，T.；安德烈兹，E.；Bifulco，P.；埃斯波西托， D. ; 中心，J。；加吉奥罗，G。可穿戴式蓝

牙分流式医疗保健监控系统。传感器2021、21、7586。PubMed

23. 哈罗德，R。蒙哥马利，战术战斗伤亡护理，快速参考指南。2017.在线可用： https://emsa。ca.政府/wp-

content/uploads/sites/71/2017/07/TCCC\_Quick\_Reference\_Guide\_2017.pdf（已于2023年4月24日访问）。

24.T.Athaya；崔，S。一篇关于估计血压波形的无创方法的综述。传感器2022、22、3953。PubMed

25.穆卡马拉，R。；斯特吉欧，G。S.；Avolio，。P.无袖带血压测量。安努。发动机的旋转比奥梅德。雕刻2022, 24, 203 –230.CrossRef 26.周，Z。B. ; Cui, T.R.；李，D.；吉安，J。M.；李，Z.；纪，S.R.。；李，X.；徐，J.D.；刘H.F.；杨，Y.；等。基于脉冲波传输时间

和脉冲到达时间的可穿戴连续血压监测装置：一篇综述。材料2023、16、2133。CrossRef

27.SondejT.；西兹科夫斯基；奥尔舍夫斯基；多布罗沃尔斯基，A。用于评估脉冲波延迟的同步多站点测量系统。双胞胎。 比奥梅德。雕刻 2019, 39, 488 –502.CrossRef

28.普莱辛格，F.；Jurco，J.；Halamek，J.；Jurak，P。信号装置：一个开放的信号处理软件平台。生理学。平均值。2016, 37, N38 – N48.CrossRef

29. 多域HMSC3是茉莉花。可在线获得：https://防御24。（已于2023年4月24日访问）。

3 0 . 谷歌远程过程调用-一个高性能的开源通用RPC框架。可在线获得： https : //grpc。io/（已于2023年4月24日访问）。 31.Pollaris，G.；萨比，M。反向分类：不仅仅是另一种方法。欧洲人J.艾默格。医学2016, 23, 240 –247.CrossRef

32.Wajtowicz, B. ; Dobrowolski, A.P. ; Tomczykiewicz, K.下降检测器使用离散小波分解和SVM分类器。地铁。平均值。西斯特。2015, 22, 303 –314.CrossRef

3 3 . 多布罗沃尔斯基。P.；Suchociki，M.；我，E。利用小波分解和SVM网络对听觉脑干反应进行分类。双胞胎。 比奥梅德。雕刻2016, 36, 427 –436.CrossRef

34.Stasiakiewicz，P.；多布罗沃尔斯基，一个。P.；塔尔戈夫斯基；Gałzka-威德里克；萨杜拉-锡克卢卡，T.；斯基科拉斯基；奥尔舍夫 斯基，R。‘as利用小波包分解和支持向量机对正常裂纹患者进行自动分类。比奥梅德。信号处理。控制2021、67、102521。CrossRef

免责声明/出版商注：所有出版物中所包含的声明、意见和数据仅为个人作者和贡献者的声明，而不是MDPI和/或编辑)的声明。MDPI和/或编 辑器对内容中提及的任何思想、方法、说明或产品对人员或财产造成的任何伤害不承担责任。

[***sensors***](https://www.mdpi.com/journal/sensors)

Article

**Decision Support System Proposal for Medical Evacuations in Military Operations †**

**Piotr Lubkowski 1,***[![](/media/202412//1733390436.2615879.png)](https://orcid.org/0000-0002-1748-5070)**, Jaroslaw Krygier 1**[![](/media/202412//1733390436.2827954.png)](https://orcid.org/0000-0002-4405-2078)**, Tadeusz Sondej 1, Andrzej P. Dobrowolski 1**[![](/media/202412//1733390436.2937078.png)](https://orcid.org/0000-0002-0593-158X)**, Lukasz Apiecionek 2,3, Wojciech Znaniecki 3 and Pawel Oskwarek 4**

![](/media/202412//1733390436.3015034.png)

check for updates

**Citation:** Lubkowski, P.; Krygier, J.;

Sondej, T.; Dobrowolski, A.P.;

Apiecionek, L.; Znaniecki, W.;

Oskwarek, P. Decision Support

System Proposal for Medical

Evacuations in Military Operations.

Sensors **2023**, 23, 5144. [https://](https://doi.org/10.3390/s23115144) [doi.org/10.3390/s23115144](https://doi.org/10.3390/s23115144)

Academic Editor: Toshiyo Tamura

Received: 28 April 2023 Revised: 20 May 2023

Accepted: 26 May 2023 Published: 28 May 2023

[![](/media/202412//1733390436.3139665.png)](https://creativecommons.org/)

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license [(https://](https://creativecommons.org/licenses/by/4.0/) [creativecommons.org/licenses/by/](https://creativecommons.org/licenses/by/4.0/) 4.0/).

1 Faculty of Electronics, Military University of Technology, Gen. SylwestraKaliskiego 2, 00-908 Warsaw, Poland;

jaroslaw.krygier@wat.edu.pl (J.K.); tadeusz.sondej@wat.edu.pl (T.S.); andrzej.dobrowolski@wat.edu.pl (A.P.D.)

2 Institute of Computer Science, Kazimierz Wielki University, Jana Karola Chodkiewicza 30, 85-064 Bydgoszcz, Poland; lapiecionek@teldat.com.pl

3 TeldatSp. z o.o. sp.k, Cicha 19, 85-650 Bydgoszcz, Poland; wznaniecki@teldat.com.pl

4 Military Institute of Medicine–National Research Institute, Szaser6w 128, 04-141 Warsaw, Poland; poskwarek@wim.mil.pl

***** Correspondence: piotr.lubkowski@wat.edu.pl; Tel.: +48-261-837-897

† This paper is supported by the project No. DOB-SZAFIR/09/B/006/01/2021 ﬁnanced by National Centre for Research and Development (NCBR)—Poland.

**Abstract:** The area of military operations is a big challenge for medical support. A particularly important factor that allows medical services to react quickly in the case of mass casualties is the ability to rapidly evacuation of wounded soldiers from a battleﬁeld. To meet this requirement, an effective medical evacuation system is essential. The paper presented the architecture of the electronically supported decision support system for medical evacuation during military operations. The system can also be used by other services such as police or ﬁre service. The system meets the requirements for tactical combat casualty care procedures and is composed of following elements: measurement subsystem, data transmission subsystem and analysis and inference subsystem. The system, based on the continuous monitoring of selected soldiers’ vital signs and biomedical signals, automatically proposes a medical segregation of wounded soldiers (medical triage). The information on the triage was visualized using the Headquarters Management System for medical personnel (ﬁrst responders, medical ofﬁcers, medical evacuation groups) and for commanders, if required. All elements of the architecture were described in the paper.

**Keywords:** enterprise architecture; decision support system; medical evacuation; vital signs measurement; biomedical signals measurement

**1. Introduction**

Medical evacuation, also abbreviated as MEDEVAC, is a set of activities provided by medical personnel to evacuate wounded soldiers or injured patients from a battleﬁeld or from a scene of accident. The ﬁrst phase of the evacuation process is devoted to a life-saving medical care of the patients at the place of accident and then to en route care provided by medical personnel in emergency medical service vehicles or aeromedical helicopters. In typical situations, the emergency service personnel reacts after receiving information on an accident coming directly form injured persons or from witnesses of the accident. The decision on medical transport of the patients to medical facilities is taken by the personnel of an ambulance (medical rescuers) after the ﬁrstaid provided at the place of accident. The situation can be more complicated if we have accidents with many casualties with different levels of severity. Often, because of lack of the resources, the medical personnel must decide which patient must be aided ﬁrstly and which can wait. In the medicine, such selection is known as a triage.

![](/media/202412//1733390436.3316746.png)

![](/media/202412//1733390436.340518.png)[MDP](https://www.mdpi.com)

Sensors **2023**, 23, 5144. <https://doi.org/10.3390/s23115144> <https://www.mdpi.com/journal/sensors>

Thus, the ﬁrst triage can be performed by the medical personnel at the scene of accident to decide about the order of patients who have to be evacuated to the hospital ﬁrstly and who can be evacuated subsequently. The triage is, in particular, required on the battleﬁeld where many wounded soldiers can suddenly appear during military action, but also at peacetime during accidents with mass casualties. To effectively support the decision on the triage, and to decrease the time of medical evacuation, medical personnel should be equipped with information on the gravity of wounds and patients conditions even before an arrival to the scene of accident. To do so, an effective decision support

system for medical evacuation (DSS-MEDEVAC) is needed. The authors of this paper propose such a system for military use, elements of which might be used by other services such as police, ﬁre service or organized rescue groups. The system was developed from scratch and it is proposed for armed forces supported by command and control information systems. The main goal of the DSS-MEDEVAC system is to monitor the health parameters of the soldiers during military actions, process these parameters and refer the decision on suggested military triage to the military personnel in case of identiﬁed life and health risks of the soldiers. Simultaneously, the information on potential medical problems can be passed to the commanders to support their decision during military operations. Based on the information from the DSS-MEDEVAC, the medical personnel that is responsible for tactical evacuation care, is able to react faster on the situation on the battleﬁeld having current information on the suggested triage and the casualties health. This information supports the medical ofﬁcers in the decision process on medical evacuations. Additionally, the information is passed to the medical evacuation vehicles or helicopters to constantly support the evacuation personnel.

The DSS-MEDEVAC system is composed of the health monitoring sensors integrated with the soldiers’ uniform and personal equipment, communication module responsible for reliable medical data transmission, decision support module responsible both for processing the data received from the sensors of each soldier and for taking the initial decision on the triage. This information supports the medical ofﬁcer, who is able to take a ﬁnal decision on medical evacuation and medical care before and during the evacuation. Additionally, the medical personnel of the evacuation vehicle can change the decision suggested by the system after the ﬁrstaid at the battleﬁeld or during the en route care. The new data are processed by the DSS-MEDEVAC system and are automatically conveyed to the deployed or ﬁxed medical facilities (decision support center, evacuation points, ﬁeld hospital). The DSS-MEDEVAC system automatically ﬁlls out an electronic TCCC card (tactical combat casualty care card) according to the information provided by its medical sensors. This electronic card continuously provides current patient data trends and can replace the paper TCCC card used at present, which is attached to wounded soldiers for evacuation and can only be updated when possible. Thus, the information on the casualties can be immediately consumed by the personnel of medical facilities to prepare required medical resources.

This paper focused on the description of the DSS-MEDEVAC system architecture, general explanation of the system components and on the presentation of the ﬁrst laboratory tests of these components. The DSS-MEDEVAC is a pre-veriﬁed system proposed for implementation in regular military units. The system is planned to be applied in Polish Armed Forces after ﬁeld tests.

We can distinguish following advantages of our system:

• It supports military medical personnel in automatic information on current health of the soldiers during military operations;

• It allows a quick reaction of medical evacuation groups in case of mass casualties on a battleﬁeld;

• It supports an automatically pre-hospital triage of wounded soldiers;

• It supports updating the electronic tactical combat casualty care cards.

The limitation of the system is that it needs a working communication system to periodically transmit the measurement data from the sensors to the inference and analysis subsystem. Thus, it cannot operate during so-called radio silence.

The paper is organized as follows. The next section describes related work regarding

similar health monitoring systems both for regular and dedicated medicine. In Section [3,](#bookmark1) the proposed DSS-MEDEVAC system architecture is clariﬁed. The elements of the system are described from Sections [4](#bookmark2)–[7.](#bookmark3) The last section concludes the paper.

**2. Related Work**

Remote health monitoring is not a new technical problem. Many solutions were proposed for wearable health monitoring sensor systems over the past several decades [[1]](#bookmark4). Additionally, many ready-for-use products can be found on the market [[2](#bookmark5)–[5]](#bookmark6). They are mainly composed of the measurement unit constantly or periodically used by the patients in their home to measure and monitorselected vital signs, data transmission unit (mostly smartphone or wi-ﬁ router) and health monitoring center, allowing visualization of threats in health alarms, measured vital signs and basic biomedical signals. These solutions are gaining importance in the age of so-called telemedicine, where a doctor can quickly react on health problems of his patients. As mentioned in the introduction, a remote health monitoring is leveraged mostly by diagnosed patients. The solutions are also equipped with the automatic call to the emergency service in case of identiﬁed critical alarms. Unfortunately, regular health monitoring systems cannot be directly used by soldiers during military operations to constantly monitor their critical vital signs.

Remote health monitoring solutions currently available on the market are based on the so-called wearable devices and have many limitations important from a military viewpoint, but importantly, they are proprietary solutions and, thus, cannot be easily integrated with military headquarters management systems. One of the interesting systems is Zephyr BioHarness [[6](#bookmark7)], which is also proposed by the manufacturer for limited military use. Unfortunately, it does not allow measuring all the vital signs we assumed to be used in our system (i.e., blood pressure, oxygen saturation), but also its communication range and technology is not acceptable for tactical operations (up to 274 m with Zephyr ECHO Gateway communication module).

An Interesting approach to remote monitoring of footballers’ health is presented in [[7](#bookmark8)], where movement responses were registered by GPS (global positioning system) units and conveyed to the analysis in order to optimize sportspersons’ performance depending on the speed intensity. To the best of our knowledge, there are no studies that used similar health monitoring methods to assess combat effectiveness or the ability to avoid injury by soldiers during military actions. We also do not assume that our system should support such a capability. Avoiding injury by the soldiers is very important, but it is not the main priority, at the theater of military operations. Therefore, activity of soldiers should mainly be optimized to reach military tasks.

The CRI (compensatory reserve index) proposed in [[8](#bookmark9)] is a novel vital sign parameter that uses features derived from machine learning of the arterial pulse wave or photoplethys- mography (PPG). In our system, we also measure the PPG signal and we have the ability to remotely send it to the analysis and inference subsystem. As indicated by the authors of [[8](#bookmark9)], the CRI parameter can be useful to identify the bleeding trauma soldiers which we will try to consider in the next version of our system.

In turn, in the article [[9](#bookmark10)], the authors explored three parameters (SI—shock index, PP—pulse pressure and ROX index) in terms of their usefulness in pathophysiologic categorization of patients in a hospital. However, these are hospital conditions, and the results of the work refer to mortality. The main role of our system is to support a decision on the soldiers’ evacuation from the battleﬁeld. Nevertheless, the above parameters were calculated on the basis of the main vital signs, i.e., heart rate, blood pressure, oxygen saturation and respiratory rate, which are also measured in our system.

A number of solutions are also dedicated speciﬁcally to support military medical care. One such solution is the NATO First Responder (NFR) application developed by the U.S. Defense Health Agency for documenting injuries and care for casualties while they are being resupplied on the battleﬁeld [[10]](#bookmark11). The application enables the preparation of reports

to be forwarded to medical services before a wounded soldier is evacuated to ﬁxed medical facilities. Communication in this case is based on the mobile networks. In the absence of access to cellular network resources, information about patients can be shared using near ﬁeld communication (NFC) tags. The aforementioned solution complies with NATO standards for the evacuation and medical protection of soldiers on the battleﬁeld. However, it is worth noting that the NFR does not support remote monitoring of wounded soldiers.

The OpenAhlta is an open source version of the U.S. Department of Defense battleﬁeld electronic medical record system (AHLTA-Theater) and operates using NFR application [[11]](#bookmark12). It uses Health Level Seven (HL7) standards for transferring clinical and administrative data. Unfortunately, it cannot be a base system for remote health monitoring of the soldiers on the battleﬁeld.

The BATDOK (battleﬁeld-assisted trauma distributed observation kit) created and owned by the U.S. Air Force Research Laboratory’s 711th Human Performance Wing is a software that can be run on a smartphone or other mobile devices to enable medics to wirelessly monitor multiple patients’ vitals at the point-of-injury [[12]](#bookmark13). The vital signs values can be delivered to the application by all available methods, starting form a measurement using accessible specialized medical devices and ﬁnishing on an observation of wounded soldiers by ﬁrst responders. This solution is conceptually very similar to our system, but it is not integrated with sensors to constantly monitor health of soldiers even before injury. It is also not integrated with a battleﬁeld management system to constantly support medical services on different levels of command.

In addition to remote health monitoring systems, some triage supporting algorithms were elaborated. Most of them are just procedures that the rescuers must apply to ef- fectively provide a medical assistance at a scene of an incident. An example of such an algorithm is START (simple triage and rapid treatment) [[13](#bookmark14)] or SALT (sort, assess, life- saving interventions, treatment and/or transport) [[14](#bookmark15)], The START algorithm supports ﬁrst responders, allowing them to triage multiple victims in a very short time, based on three observations,i.e., respiration, perfusion and mental status. Responders assign patients to four categories: deceased, immediate, delayed and minor. Triage tags are used (mostly physical marking) to distinguish patients sorted to separate categories. Similarly to the START, the SALT is a four-step process helping ﬁrst responders to manage mass casualty incidents. It also proposes the tags to segregate the patients (dead, immediate, expectant, delayed and minimal). Both START and SALT are relatively simple triage algorithm based on vary basic observations; thus, they often are used at the ﬁrst stage of segregation pro- cess. More advanced systems such as TEWS (triage early warning system) [[15](#bookmark16)] or NEWS2 (national early warning score 2) [[16](#bookmark17)] require more measurements (i.e., blood pressure, heart rate, oxygen saturation and others) to triage casualties.

The authors of [[17](#bookmark18)], in their paper, discussed one of the triage methods that can be successfully applied in regular emergency services. This method refers to the triage of the patients who subsequently required hospital admission and who were likely to die within 24 h. Since available criteria pointed out in the cited paper can be used for selecting soldiers who can quicker be recovered and returned to military actions, we propose nearly the same approach in our system for so-called reverse triage, but we should remember that tactical operations often preclude rapid medical assistance and evacuation, where rescuers often cannot immediately reach the point of injury and evacuation process is delayed (the military priorities are more important than medical evacuation of wounded soldiers during military operations). Having the constant monitoring system that is based not only on the measurement of respiratory rate or oxygen saturation, we can offer the military medical personnel means for fast reaction even before arriving to the point of injury or even during lack of medical assistance. The medical evacuation personnel can remotely instruct (using military radios) ﬁrst responders (neighboring soldiers which can help wounded soldiers) based on the information coming from our system. Additionally, the basic vital signs that we monitor only suggest the initial triage, which must be conﬁrmed by ﬁrst responders (commander or the soldiers which are able to prepare and send requests for medical

evacuation), where easily available criteria pointed out in the cited paper are applied. In addition to the information on the initial triage, our system allows sending certain biomedical signals that are generated at the request of medical personnel alerted by initial triage. All the above actions can be initiated automatically owing to constant monitoring of selected vital signs. However, it should be noted that there are also some limitations of our system. This situation will take place during radio silence (speciﬁcmilitary situations where radios cannot transmit data for security reasons), but this limitation is valid for all remote monitoring systems.

Taking into account the mentioned triage support systems, the authors prepared a set of capabilities which the DSS-MEDEVAC system have to achieve. The required capabilities of the DSS-MEDEVAC are described in Section [3.](#bookmark1)

**3. Proposed Architecture and Requirements of the Decision Support System for** **Medical Evacuation (DSS-MEDEVAC)**

A system that should support medical personnel during military operations will differ from typical healthcare telemonitoring systems, although the goal of such systems is similar in both cases; they must measure the relevant vital signs of the patients/soldiers/ﬁreﬁghters, and transfer the data to the health monitoring center in order to react faster on critical health events. However, in a military, tactical environment (on the battleﬁeld),the main priority is to achieve the military goal. Therefore, the rapid recovery of the soldiers to use them again in combat is a crucial task for medical personnel. To cope with this task, a reverse triage is often used during pre-hospital medical care.

To help military medical personnel to make decisions on fast evacuation of the wounded soldiers from a battleﬁeld, constant health monitoring is required, not focusing on speciﬁc illness of a soldier (but taking into account personal health parameters). Such constant monitoring requires not only a reliable measurement of human vital signs, but also should take into account the speciﬁc battleﬁeld conditions (the soldiers can run, drop, can be overstressed, and eventually can be slightly or severely wounded, or can suffer death on a battleﬁeld). Taking such conditions into account, the soldiers’ health monitoring system will differ from the health monitoring solutions available on the civilian market. The authors prepared the architecture of such system, focusing on supporting medical personnel on a battleﬁeld, especially in terms of faster triage before and during medical evacuation process. The abbreviations used during system description are collected in Table [1.](#bookmark19)

**Table 1.** List of abbreviations used during DSS-MEDEVAC system description.

**Abbreviation**

**Meaning**

AAMB

Air Ambulance

AFE

Analog Front End

AIC

Analysis and Inference Capability

AIS

Analysis and Inference Subsystem

AMB

Ambulance

BAN

Body Area bio-sensors’ Network

BMMC

Battleﬁeld Medical Monitoring Centre

BMS

Battleﬁeld Management System

BodyPos

Body Position

C3IS

Command, Control, Communication, Intelligence and Surveillance

CN

Commander Node

CPU

Central Processing Unit

DMN

Decision Making Node

DTC

Data Transmission Capability

DTS

Data Transmission Subsystem

ECG

Electrocardiogram

**Table 1.** Cont.

**Abbreviation**

**Meaning**

gRPC

Google Remote Procedure Calls

HMS

Headquarters Management System

HR

Heart Rate

IP

Internet Protocol

MDTP

DSS-MEDEVAC Data Transmission Protocol

MEG

Medical Evacuation Groups

MEN

Medical Evacuation Node

MMC

Medical Monitoring Center

MPA

Mean Physical Activity

MSAC

Medical Situation Awareness Capability

MTF

Medical Treatment Facility

NAFv4

NATO Architecture Framework version 4

NVIC

Nested Vectored Interrupt Controller

PAS

Physical Activity Signal

PAT

Pulse Arrival Time

POI

Point of injury

PPG

Photoplethysmogram

PTT

Pulse Transit Time

RES

Respiratory Rate Signal

RR

Respiratory Rate

SBC

Single Board Computer

SN

Soldier Node

SpO2, SpO2

Oxygen Saturation

SVM

Support Vector Machine

TCCC

Tactical Combat Casualty Care

UART

Universal Asynchronous Receiver-Transmitter

USB

Universal Serial Bus

SBP

Systolic Blood Pressure

VCP

Virtual COM Port

3.1. System Requirements

The architecture of DSS-MEDEVAC system dedicated for military operations was prepared based on the NATO Architecture Framework version 4 (NAFv4) [[18](#bookmark21)], which is a standard for developing and describing architectures for both military and business use. The NAFv4 supports preparation of different viewpoints of the developed system, including concept of the system (C group of viewpoints), service speciﬁcations (S group of viewpoints), logical speciﬁcations (L group of viewpoints), physical resource speciﬁcations (P group of viewpoints) and architecture foundation (A group of viewpoints).

In this section, the main viewpoints of the DSS-MEDEVAC system are described,i.e., Capability Taxonomy (C1),Enterprise Vision (C2),Capability Dependencies (C3), Service Taxonomy (S1), Node Types (L1) and Logical Scenario (L2).

To reﬂect the goal of the DSS-MEDEVAC system, the main system capabilities are deﬁned. Figure [1](#bookmark22)shows proposed the simpliﬁed system Capability Taxonomy (C1) and Capability Dependencies (C3) viewpoints.

It was assumed that the system must present to the medical personnel and selected commanders the current information on the soldier’s vital signs and initial decision on the triage, in the case of identiﬁed (detected) life or health threats (medical situation awareness capability—MSAC). Additionally, the system must present current location of each soldier to help organize the medical evacuation. The system must be capable of measuring, registering and processing the required human vital signs (vital signs monitoring capability) and reliably send the results of the measures (data transmission capability—DTC) to the analysis and inference module. Then, the system must effectively process the measured data to propose initial decision of the triage (analysis and inference capability—AIC).

V

<<Capability>>

Electronic

TCCC Card

<<Capability>> Visualisation

<

>

<<Capability>>

MEDEVAC

Decision Support,

Medical Triage

<<Capability>>

Medical Situation

Awareness

<<Capability>>

Calibration and

Personal

Reference Data

<

<<Capability>>

Analysis and

Inference

<<Capability>> Vital Signs Monitoring

<<Capability>>

Biomedical

signals

Monitoring

>

<

<<Capability>>

Data

Transmission

<<Capability>>

HR

Measurement

<<Capability>>

PAS

Measurement

<<Capability>>

ECG

Measurement

>

V

<

<<Capability>>

RR

Measurement

<<Capability>>

BodyPos

Measurement

<<Capability>>

PPG

Measurement

<<Capability>>

SpO2

Measurement

<<Capability>>

SBP

Measurement

![](/media/202412//1733390436.3882089.png)

<<Capability>>

MPA

Measurement

<<Capability>>

RES

Measurement

**Figure 1.** DSS-MEDEVAC capability taxonomy and capability dependencies.

After detailed analysis, it was decided that the main goal of the system can be achieved by supplying the AIC by following human vital signs: heart rate (HR), respiratory rate (RR), oxygen saturation (SpO2 ), mean physical activity (MPA),body position (BodyPos) and systolic blood pressure (SBP).

The set of vital signs was selected on the basis of four factors,i.e.:

• Related work analysis on proposed solutions supporting the triage during mass casualty incidents (for example: [[19,](#bookmark23)[20](#bookmark24)]);

• Related work analysis on supporting remote health monitoring during disaster- induced mass casualty incidents by Internet of Thigs (IoT) solutions (for example: [[21](#bookmark25)]);

• Analysis of similar solutions dedicated for the military (for example: [[10](#bookmark11)–[12](#bookmark13)]);

• Personal experience of ﬁrst responders and pre-hospital medical personnel on medical evacuation during military operations.

The authors of [[21](#bookmark25)] proposed a real-time monitoring electronic triage tag system for improving survival rate in disaster-induced mass casualty incidents—the goals of this system were, therefore, similar to the goalsof our system. They also proposed vital signs that can be simply measurable during pre-hospital triage,i.e., body temperature, heart rate, blood pressure, respiratory rate, oxygen saturation, capillary reﬁll, consciousness and electrocardiogram. The authors of [[21](#bookmark25)] also argued that vital signs such as heart rate, blood pressure, respiratory rate and electrocardiogram are considered to be the most important factors by ﬁeld rescuers. Since the triage in our system is based on the simple triage and rapid treatment (START) procedures, we also proposed a set of vital signs that can be constantly, autonomously, non-invasively and, importantly, simply measurable

during military operations (for example,SpO2 and HR), even without a specialized medical equipment (for example, RR).

The selection of the vital signs for our system also resulted from the analysis of other triage systems (procedures). Most triage systems are based on a simple test of consciousness using ACVPU (Alert, Confusion, Voice, Pain, Unresponsive) scale, presence of a pulse on the radial artery and the number of respirations, but in more advanced triage systems, SpO2, SBP, HR are also measured. In [[19](#bookmark23)], an extensive analysis of the currently functioning triage systems is shown. The authors of [[20](#bookmark24)] also performed an analysis on the triage systems and useful vital signs relating to the mass-casualty bioterrorism. The vital signs we selected for our system are often the main variables taken into account in analyzed triage systems. The analysis led us to the conclusion that we need a minimal set of vital signs that is able to support rapid remote inference about a soldier’s health status, taking into account strong military restrictions and procedures.

The AIC should also be supplied by the calibration and reference data, speciﬁc for each soldier (calibration and personal reference data-loading capability). Additionally, the system must enable the medical ofﬁcers or rescuers to remotely observe required biomedical signals of a selected soldier what can help them to prepare in advance required medical equipment and medications, and to update the decision on the triage, if required (biomedical signals monitoring capability). Following biomedical signals of the soldier being assumed to be monitored remotely on demand: electrocardiogram (ECG), photo- plethysmogram (PPG), physical activity signal (PAS) and respiratory rate signal (RES). The registered vital signs of the wounded soldiers must be automatically passed to the electronic tactical combat casualty care card, which must be immediately available in the system for the medical personnel involved in the soldiers’ treatment (Electronic TCCC Card Filling Capability). Both the TCCC card and the initial decision on the triage must be immediately available for viewing in ‘realtime’ in the headquarters management system (HMS) used by the Armed Forces during military operations (visualization capability).

3.2. Realization of System Requirements

A service taxonomy viewpoint (S1) is shown in Figure [2.](#bookmark26) The services offered by the DSS-MEDEVAC system reﬂects the capabilities that are required (shown in Figure [1)](#bookmark22). The main service group is responsible for visualization of the information that is concerned with the triage decision, soldier’s localization, vital signs, biomedical signals and electronic TCCC card. The important service required by the system is also the data transmission service, which ensures reliable data distribution over all the players in the system.

<<ServiceSpecification>>

Vital Parameters

Visualisation

Service

<<ServiceSpecification>>

Biomedical signals

Visualisation

Service

<<ServiceSpecification>>

Medical Triage

Visualisation

Service

<<ServiceSpecification>>

Localisation

Visualisation

Service

<<ServiceSpecification>>

Analysis and

Inference Service

<

<<ServiceSpecification>>

e-TCCC Card

Visualisation

Service

<<ServiceSpecification>>

HR, RR, SpO2, SBP,

MPA, BodyPos

Measurement Service

<<ServiceSpecification>>

Data Transmission

Service

<<ServiceSpecification>>

ECG, PPG, RES,

PAS Measurement

Service

**Figure 2.** DSS-MEDEVAC service taxonomy viewpoint.

The operational viewpoint of the medical evacuation actions where the DSS-MEDEVAC system is planned to be employed is shown in Figure [3.](#bookmark27) Wounded soldiers are ﬁrstly evac- uated to a secure area at the battleﬁeld (if it is possible) to give them the ﬁrstaid. Let us name this area the point of injury, where the ﬁrst professional medical assistance can be delivered by the medical personnel (military ambulance—AMB or air ambulance—AAMB) after reaching this area. The sensors of DSS-MEDEVAC system should ﬁrstly react on the life problems of the casualties; thus, medical personnel can be supported by important life parameters at the early stage of the evacuation process. The initial triage is also suggested immediately; thus, the required medical resources can be involved in the medical evac- uation process. Additionally, current biomedical signal of the wounded soldiers can be observed before the evacuation and during the en route care (supporting the professional medical instruments). The upper levels of the military evacuation pointed out in Figure [3](#bookmark27) can exploit both the electronic TCCC card and the information on the initial and updated triage and the current conditions of wounded soldiers. The operational viewpoint shows the main players involved in the DSS-MEDEVAC systems.

![](/media/202412//1733390436.4146943.png)![](/media/202412//1733390436.4255965.png)![](/media/202412//1733390436.434585.png)

AAMB

MTF

Legend:

4

1 Point of injury

(POI)

3

![](/media/202412//1733390436.4585896.png)![](/media/202412//1733390436.478625.png)![](/media/202412//1733390436.4879913.png)

![](/media/202412//1733390436.497541.png) Batalion Aid

2

AAMB

AMB

Station

![](/media/202412//1733390436.5074944.png)

![](/media/202412//1733390436.51928.png)![](/media/202412//1733390436.528998.png) ![](/media/202412//1733390436.538952.png)

3

Brigade Medical Point

AMB

2

AAMB

4

2

![](/media/202412//1733390436.5504105.png)MTF – Medical

![](/media/202412//1733390436.5591595.png)![](/media/202412//1733390436.5680027.jpeg)Treatment Facility

![](/media/202412//1733390436.5768535.png)AMB

![](/media/202412//1733390436.588336.png)AMB - Ambulance

AMB

1 POI

AMB

![](/media/202412//1733390436.5982165.png)AMB

AAMB – Air Ambulance

1 POI

1

POI

**Figure 3.** DSS-MEDEVAC operational (enterprise) viewpoint.

Figure [4](#bookmark28)clariﬁes the DSS-MEDEVAC system general viewpoint. At the single soldier level, a set of bio-sensors have to be employed to measure both the vital signs and biomed- ical signals, which are important for the initial triage preparation and from all medical evacuation process point of view (indicated and explained in the capability viewpoint).

The general information on the soldier’s life-condition (initial and ﬁnal triage) is visu- alized by the battleﬁeld management system (BMS) to the direct commander, supporting the current situation awareness. The set of vital signs are constantly sent using the pre- pared communication protocol via the communication network to the medical monitoring center (MMC), where the data are analyzed to propose the initial and ﬁnal decision on the triage. This information is then passed to the visualization module available to the medical personnel of theMCC, which can take a ﬁnal decision on medical evacuation. The medical personnel can also observe the required biomedical signals of the wounded soldiers to additionally support the decision on evacuation or to advice remotely on the ﬁrstaid. The same information appears in the visualization module of the medical evacuation groups (MEG)—ambulances. The medical personnel of the MEG can change the decision on the triage after direct medical actions at the point of injury. Initially ﬁlled in the electronic

TCCC card, associated with each wounded soldier, is also updated by the rescuers and is immediately available for medical personnel at each level of the evacuation facilities.

![](/media/202412//1733390436.6138546.png)

**Figure 4.** DSS-MEDEVAC system general viewpoint.

The logical viewpoint of the DSS-MEDEVAC system is presented in Figure [5.](#bookmark29) Three main nodes are distinguished in the system: Soldier Node (SN), Commander Node (CN), Decision Making Node (DMN) and Medical Evacuation Node (MEN).

The SN is composed of a body area bio-sensors’ network (BAN), a single board computer (SBC), a personal terminal and a personal radio. The BAN is a network integrated with personal equipment of the soldier (including military uniform, helmet and underwear). The SBC is designed to perform initial processing of biomedical signals and real-time computing of vital signs and to refer the processed data to the personal terminal via a developed communication protocol. The personal terminal is a tablet-based military computer, which constitute the gateway to the tactical data transmission system. It is also assumed that the BAN and, particularly, the SBC can monitor and store the history of the vital signs of the soldier even if the other equipment is inaccessible. These values can be accessible later on if required. The tactical data transmission system is a part of the BMS and HMS systems. Its main role is to reliably distribute the medical data over all elements of the DSS-MEDEVAC system using tactical radio networks (personal radios, vehicular radio equipment, deployed and ﬁxed communication infrastructure).

![](/media/202412//1733390436.6304302.png)

**Figure 5.** DSS-MEDEVAC logical viewpoint—nodes types viewpoint.

The CN role is to collect the medical data transmitted by subordinates and to send them to the DMN. Based on the data received from the CNs, the DMN node performs the advanced analysis and inference and suggests the decision on the triage. The decision is passed to the visualization modules located in commander nodes, medical evacuation nodes and other medical facilities involved in the medical evacuation actions. The medical data are distributed between the nodes of the DSS-MEDEVAC using the tactical data transmission system. The information on the triage, vital signs and biomedical signals is

visualized both to the medical personnel and the commanders with the use of BMS and HMS systems. The medical evacuation node (ambulance, aeromedical helicopter), receives the current decision on the triage and initially ﬁlled in the electronic TCCC card to support the rescuers. The MEN is also the source of the information on the ﬁnal decisions which are passed to the DMN and distributed over the system.

The main subsystems of the DSS-MEDEVAC system, which reﬂect the components of the logical viewpoint presented in Figure [5](#bookmark29), are described in the next sections.

**4. Measurement Subsystem**

Using different kind of sensors for measurement of the biomedical signals and vital signs in a relatively static body position is well described in the literature. In the DSS- MEDEVAC system, we assume difﬁcult measurement conditions, where some sensors can be destroyed or disconnected and the analysis of the received data should predict such situations. Moreover, we should assume that commonly known wireless transmission techniques dedicated for sensor networks (i.e., Bluetooth, ZigBee, ANT or Near Field Communications) cannot be applied during military operations, especially because of intentional jamming of radio transmissions. All these considerations lead to the decision on a set of health parameters that should be monitored, initially processed and sent to the analysis and inference subsystem, as well as to the decision on the communication technique that prevents data transmission from jamming. Additionally, wearable sensors should be integrated with a personal soldiers’ equipment (uniform, underwear, communication equipment). A simpliﬁed architecture of the measurement subsystem is shown in Figure [6.](#bookmark30)

![](/media/202412//1733390436.6449034.png)1

1 - Forehead sensors: PPG, ECG (optional)

Personal Radio

2 – Chest sensors:

Breath Sensors, ECG, Accelerometer

2

3 – Wrist sensors (optional) PPG, ECG, Accelerometer

3

Personal Terminal

SBC – Single Board Computer

**Figure 6.** General architecture of the measurement subsystem of the DSS-MEDEVAC.

Wearable sensors, constructed by the authors of this paper, are located on the forehead, chest and wrist (optional sensor) of a soldier. They are integrated with a helmet and with an uniform. A single board computer (SBC) based on high-performance arm cortex-M7 microcontroller with low power consumption is responsible for data acquisition from sensors, initial processing of the measured signals, real-time computing of vital signs and communication with personal terminal. To ensure jamming protection, data are transferred to the SBC using wired communication, where wires are to be integrated with an uniform.

Both the SBC and sensors are powered by a personal terminal (as a main energy source) or by an internal SBC battery (in a stand-by or emergency mode). It is assumed that the personal terminal is able to deliver required energy to the measurement subsystem during the military mission. The data are sent to other parts of the system via a personal radio which ensures secure and robust tactical communication (even under hostile jamming).

In general, a large set of biomedical signals and vital signs can be measured to support remote health monitoring [[2,](#bookmark5)[22]](#bookmark31). Nevertheless, we do not have to use all of them to support a decision on the triage in military operations (or in similar actions). Taking into account the required capabilities of the DSS-MEDEVAC system, speciﬁc military conditions, and after deep analysis of the military health care and evacuation procedures [[23](#bookmark32)] during military actions, we decided to measure biomedical signals and vital signs pointed out in Figure [1.](#bookmark22) These signs are supported by information on soldier’s motion and body position (by using of an accelerometer) as shown in Figure [6.](#bookmark30) On the current stage of the system development, authors prepared a testbed with all required sensors, where each sensor is composed of two main components: CPU (central processing unit) and AFE (analog frontend) modules, shown in Figure [7.](#bookmark33) A sensor is able to communicate with the SBC via two-wires UART (universal asynchronous receiver-transmitter) interface. Such a modular architecture of the sensors allows using dedicated AFE for each sensor with the same CPU module.

UART

interface to

SBC

Powering

Sensors’ architecture

CPU module

Board to board interface

AFE module

**Figure 7.** DSS-MEDEVAC sensors’ architecture.

As shown in Figure [6](#bookmark30), measurement subsystem of the DSS-MEDEVAC contains three sensors measuring biomedical signals. These signals are integrated in the SBC computer; therefore, it is important to ensure synchronization of the measurement of these signals. This is especially important when calculating the SBP parameter. The SBP will be measured continuously, cufﬂess and non-invasive, based on the propagation time (PAT—Pulse Ar- rival Time or PTT—Pulse Transit Time) of the pulse wave calculated from the ECG/PPG signals [[24](#bookmark34)–[27]](#bookmark35). In our measurement subsystem, shown in Figure [6, we proposed a dis](#bookmark30)- tributed, wired system of sensors synchronized by the SBC. To control the sensors, the SBC periodically sends requests to each sensor. Each of the sensors must respond to the request (i.e., it sends signals samples) in a ﬁxed time, less than the highest sampling period of the requested signal (e.g., for ECG the response time must be less than 4 ms). In our sensors, we used STM32L5 series of the microcontrollers that have nested vectored interrupt controller (NVIC), allowing preparing up to eight interrupt priorities. The requests received by the sensors from the SBC (i.e., interrupt of receiving a byte from the UART interface) have an appropriate high priority. Thanks to this, the sensor’s response (i.e., starts data transmission via UART) time is below 100 μs. Figure [8](#bookmark36)shows that developed AFE modules are able to measure and send the ECG and PPG signals, which are then sent to the SBC.

![](/media/202412//1733390436.6634552.jpeg)

**Figure 8.** Example signals plot measured by the AFE modules of the sensors (“,” means decimal separator).

The SignalPlant [[28](#bookmark37)] software was used to display the data in Figure [8.](#bookmark36) It facilitates viewing multiple signals with the same sampling rate and has many useful functions, including data processing.

**5. Data Transmission Subsystem**

A role of the data transmission subsystem (DTS) is to provide the data transmission service (pointed out in Figure [2](#bookmark26)) to the DSS-MEDEVAC system. The DTS integrates SN, CN, DMN and MED nodes, as shown in Figure [9.](#bookmark38)

CN

SN

MED

DMN

DTS

**Figure 9.** Data Transmission Subsystem as an integrator of the DSS-MEDEVAC nodes.

Reliable medical data distribution between the nodes are ensured by the headquarter management system (HMS), initially developed to support command, control, commu- nication, intelligence and surveillance (C3IS) capability during military operations [[29]](#bookmark39). The HMS C3IS system in its basic form enables commanders to make decisions in a timely manner, by acquiring information reported, creating a common picture of the operational situation and exchanging operational information. Thus, the HMS can also be powered by the information form DSS-MEDEVAC to support medical personnel at each level of the military medical care, including military medical evacuation activities (as shown in Figure [4)](#bookmark28). A core of the DTS is a tactical communication network based on the Internet

protocol (IP). The HMS working on the top of the IP-based communication network ensures reliable database updating and replication. In this section, we focused on one part of the DTS,i.e., on the method of data delivery to the HMS.

Figure [10](#bookmark40)shows a communication component of the soldier node.

Single Board Computer (SBC)

![](/media/202412//1733390436.6784267.png)

![](/media/202412//1733390436.6881764.png)

VCP

interface

MDTP (Device Mode)

MDTP (Host Mode)

USB

(Device Mode)

USB (Host Mode)

Personal Terminal

![](/media/202412//1733390436.701028.png)

gRPC Server

Iv gRPC/ IP interface

VCP

interface

HMS C3IS Integrator

gRPC Client

IP

UART interface

USB

interface

Sensor

interface

Personal Radio

**Figure 10.** Communication component of the soldier node.

As was already mentioned, the sensors are controlled by the SBC via the UART in- terface. Next, the data are transferred to the personal terminal using the universal serial bus (USB) interface, where the SBC runs the USB driver with a device mode, while the personal terminal uses the USB driver with a host mode. To coordinate the data transfer via the USB interface, a MDTP protocol (DSS-MEDEVAC data transmission protocol) was elaborated and implemented, both in the SBC (MDTP device mode) and in the terminal (MDTP host mode). Virtual COM Ports (VCP) are used by the MDTP drivers to commu- nicate with the USB drivers. The data are sent to/from the HMS C3IS system using the Google Remote Procedure Calls (gRPC) [[30](#bookmark41)], which are transferred by IP packets between the gRPC client and server applications. The gRPC client is a part of the integrator which distributes the data over the DSS-MEDEVAC nodes. The Iv interface is responsible for inter-process communication.

**6. Analysis and Inference Subsystem**

In this section,just general information on the triage algorithm is provided, which can clarify the role of the analysis and inference subsystem (AIS) in the DSS-MEDEVAC system. A heart of the AIS is a decision algorithm segregating wounded soldiers by assigning them following colors: green, yellow, red and black. The ﬁrst three colors reﬂect the requirements on the triage deﬁned in the START (simple triage and rapid treatment) procedures. The last one was added to indicate communication problems.

We also proposed the appropriate values of the chance of soldiers’ survival metric based on the assessment of practical cases resulting from experiences of the medical evacuation personnel. These values will be optimized after the implementation of the system in the armed forces based on the lessons learned. At the moment, we veriﬁed our assumptions in the initial ﬁeld tests, which conﬁrmed that all elements of the system can be integrated with the battleﬁeld management system, but for obvious reasons, we were unable to verify it on a real battleﬁeld with a large number of wounded soldiers. Thus, we assumed that system is able to learn and modify its initial attributes.

The green color means that chance of survival of a soldier is equal to 100%. It means that a soldier can indeed be slightly wounded, but no medical assistance is needed. The yellow color is assigned to wounded soldiers who should be evacuated but with the second priority. The red color is reserved for soldiers requiring immediate evacuation (with the ﬁrst priority). If a heart rate signal (HR) is not received from a soldier’s measurement devices, the black color is assigned. It can mean that communication problem accrued or a soldier is dead (this color requires additional veriﬁcation by medical personnel or other soldiers). In addition to the colors deﬁned above, the blue color can also be assigned to the seriously wounded soldiers with no (or minimal) chance of survival. This color can only be assigned manually by medical personnel or rescuers from the medical evacuation group,

after veriﬁcation of the soldiers’ health at the point of injury (typically on the battleﬁeld). According to miliary procedures, such soldiers will be evacuated with the last priority (known as revers triage).

The system requirements described in Section [3.1](#bookmark20)mandate that AIS should be supplied by four main vital signs to make a decision on the triage,i.e., heart rate (HR), respiratory rate (RR), systolic blood pressure (SBP) and peripheral oxygen saturation (SpO2 ). The algorithm can also be supported by information on the physical activity and body position (estimated by accelerometers). The highest priority in soldiers health assessment, in the context of medical segregation, has the HR. The RR is a parameter that rapidly reacts on the changes in human health condition. The SBP is a vital sign resulting from both HR and RR. The value of this parameter reacts slower on deteriorating patient condition. The SpO2 is less credible than previous parameters, since it depends on many factors including those not directly connected with threat to life and health; thus, it should betaken into account in last order to prepare a decision on the triage.

The AIS does not require the biomedical signals which are measured by the measure- ment subsystem. They are sent on demand directly to the medical personnel located in the medical monitoring center or to the rescuers approaching to the scene of accident or battleﬁeld. Thus, the biomedical signals support the ﬁrstaid and the ﬁnal decision on the triage.

The main triage algorithm was clariﬁed in a pseudocode shown in Algorithm 1. According to it, the (green) color is assigned to the soldiers if all measured parameters are in the ranges of the personalized reference values. The (yellow) color is assigned if a value of at least one parameter is outside the reference range but it does not exceed the critical value (which indicates threat to health). One of the rules of the algorithm is that if value of more important parameter enforced the yellow or red color, the triage color cannot be changed down (i.e., to green or yellow color, respectively). After the analysis, the most important vital sign is RR, then HR, SBP and, ﬁnally, SpO2 . Thus, in such an order, the signs were analyzed in Algorithm 1 (RR from line 1, HR from line 8, SBP from line 23 and SpO2 from line 38).

The main algorithm is additionally supplemented by a set of algorithms prepared to assess a chance of survival of wounded soldiers and reacting on lack of some parameters from the sensors (what can be often observed during military actions).

On the basis of the main algorithm, an auxiliary algorithm was deﬁned that deter- mines the function of survival chances. The need to deﬁne such a function results from the expectations related to the introduction of software supporting triage. The order of evacuation results from the assigned status, marked with the appropriate color, but within the same color, the value of the function of survival chances maybe decisive.

In the case of the so-called tactical reverse [[31](#bookmark42)]) triage—as part of TCCC (tactical combat casualty care)—depending on the tactical situation, it is used when the priority is to regain the ability to perform tactical operations as soon as possible, so the lightest wounded will be rescued ﬁrst; thus, they will be able to return to a tactical action as soon as possible, to recreate combat capability. This is a speciﬁc situation in which the least injured are saved in order to be able to use them again in combat. In this situation, the value of the function of survival chances will also be very helpful. In general, for the rescuer, the more information on the input, the better, and the function of survival chances—colloquially speaking—allows you to assess which of the red ones is potentially “more” red than the rest, which can improve the actions of medics in terms of prioritizing evacuation.

In order to deﬁne the chance of survival function, 7200 cases were generated that covered the entire four-dimensional parameter space quite well. These cases were used to train two non-linear SVM (support vector machine) networks [[32](#bookmark43)–[34](#bookmark44)] and the ﬁnal result was a function that assigns the red class a range of 1–50%, the yellow class 51–99% and the green class 100%.

**Algorithm 1** The main triage algorithm implemented in the Analysis and Inference Subsystem.

**Input:**

Measured values of RR, HR, SBP and SpO2.

Personalized reference and critical values of the vital signs for each soldier.

Typical reference values:

RRref: 9–20/min

HRref: 50–110/min

SBPref: 100–180 mmHg SpO2ref: >= 94%

RRctitical,HRcritical, SBPcritical,SpO2Critical

**Output:**

Triage

**Steps:**

1. Read RR

2. if RR == RRref then

3. | triage <- (Green)

4. else if RR != RRctitical

5. | triage <- (Yellow)

6. else

7. triage <- (Red)

8. Read HR

9. if triage == (Green)

10. | if HR == HRref then

11. | | triage <- (Green)

12. | elseif HR != HRcritical

13. | | triage <- (Yellow)

14. | else

15. | triage <- (Red)

16. elseif triage == (Yellow) then

17. | if HR == HRref or HR != HRcritical then

18. | | triage <- (Yellow)

19. | else

20. | triage <- (Red)

21. else

22. triage <- (Red)

23. Read SBP

24. if triage == (Green)

25. | if SBP == SBPref then

26. | | triage <- (Green)

27. | elseif SBP != SBPcritical

28. | | triage <- (Yellow)

29. | else

30. | triage <- (Red)

31. elseif triage == (Yellow) then

32. | if SBP == SBPref or SBP != SBPcritical then

33. | | triage <- (Yellow)

34. | else

35. | triage <- (Red)

36. else

37. triage <- (Red)

38. Read SpO2

39. if triage == (Green)

40. | if SpO2 == SpO2ref then

41. | | triage <- (Green)

42. | elseif SpO2 != SpO2critical

43. | | triage <- (Yellow)

44. | else

45. | triage <- (Red)

46. elseif triage == (Yellow) then

47. | if SpO2 == SpO2ref or SpO2 != SpO2critical then

48. | | triage <- (Yellow)

49. | else

50. | triage <- (Red)

51. else

52. triage <- (Red)

53. End of the algorithm

To train the SVM network, we used the data generated by a high-ﬁdelity simulator of vital signs of injured patients. The simulator was validated by health parameters measured in trauma patients. This decision resulted from the lack of a sufﬁcient number of measurements carried out on real patients and were only used to verify our algorithms. However, it should be emphasized that SVM network will be trained during the ﬁnal operation of the system to optimize its behavior. Red, yellow and green classes ranges should be considered as initial values which are planned to be optimized based on the military lessons learned.

The graph of chance of survival function for selected parameters is shown in Figure [11.](#bookmark45) It should be noted that the presented function will also be optimized based on the military lessons learned. We can expect that, in the initial stage of our system’s operation, there may be false alerts, but using learning capability, it should adapt its performance.

![](/media/202412//1733390436.7318695.png)

**Figure 11.** Graph of the chance of survival function for selected values of vital signs.

It is not uncommon for one or more sensors to be cutoff on the battleﬁeld. This may be due to signal cut-off or strong interference preventing the correct determination of the parameter. As a consequence, the following variants of the system operation were used:

1. All four parameters are available. The full algorithm is used and—only in this case—the value of the chance of survival function is calculated. The reliability is 100%.

2. If the SpO2 sensor is disconnected, an algorithm using only HR, RR and SBP is applied. The reliability is 90%.

3. If the SBP sensor is disconnected, an algorithm using only HR, RR and SpO2 is applied. The reliability is 80%.

4. If the SpO2 and SBP sensors are disconnected, an algorithm using only HR and RR is applied. The reliability is 70%.

5. If the RR sensor is disconnected, an algorithm using only HR, SBP and SpO2 is applied. The reliability is 80%.

6. If the RR andSpO2 sensors are disconnected, an algorithm that uses only HR and SBP is used. The re-liability is 70%.

7. If the RR and SBP sensors are disconnected, an algorithm that uses only HR and SpO2 is used. The reliability is 60%.

8. If the RR, SpO2, and SBP sensors are disconnected, the HR-only algorithm is used. The reliability is 50%.

9. In the absence of a signal from the HR sensor, the system does not work (regardless of the others). The soldier is marked as “Black”, but it is necessary to conﬁrmit by the medic at the scene.

A general architecture of the AIS is shown in Figure [12.](#bookmark46)

Analysis and

Inference Server

HMS C3IS system

![](/media/202412//1733390436.7456355.png) HMS C3IS

Integrator

AIS algorithms

gRPC Server

gRPC Client

DB

Iv

gRPC/ IP interface

**Figure 12.** Analysis and Inference Subsystem architecture.

The AIS algorithms are implemented in the Analysis and Inference Server which communicates with the HMS C3IS system via the IP interface using the Google Remote Procedure Calls. The AIS algorithms are supplied by vital signs of each soldier. To coordinate current calculations with historical data, both values of the parameters and the results are stored in a database (DB). The information on the triage is sent to the HMS C3IS system each time the triage color is changed. The HMS C3IS system is responsible for visualization of the triage to the commander (in limited form, required for command and control), medical personnel or rescuers.

An example window with suggested decision on the triage visualization for selected soldier is shown in Figure [13.](#bookmark47)

![](/media/202412//1733390436.7573352.jpeg)

**Figure 13.** Example window with suggested triage color generated by AIS algorithm.

Based on the parameters delivered to the AIS, the algorithms suggested that soldier should be tagged by the (yellow) color. All the vital signs are accessible to the AIS; thus, the reliability of the results is 100%. The SVM network suggests that chance of survival of the soldier is 74%. The ﬁnal decision on the medical evacuation and the triage is now taken by medical personnel, who can additionally check the required biomedical signals of the soldier.

**7. Visualization Module for Medical Personnel and Commanders**

In order to visualize the collected data for operators of Medical Support Groups

(MSG)/Battleﬁeld Medical Monitoring Centre (BMMC), the dedicated portal was designed and implemented as a part of the Polish headquarter management system (HMS C3IS JASMINE) [[29]](#bookmark39). The portal provides access to information about soldiers monitored by the system. The MSG Portal view adapted to the T5” TRYTON Tactical Computer Terminal is shown in Figure [14.](#bookmark48) The main view presents basic data allowing medical personnel or commander to be oriented in a battleﬁeld situation: a list of soldiers, the latest outcome of the triage and information about data updates.

![](/media/202412//1733390436.7720363.jpeg)

**Figure 14.** MSG Portal: view adapted to the T5” TRYTON Tactical Computer Terminal.

Using the software, the MSG operator has access to the detailed data of each monitored soldier. The portal provides obtaining the following data:

• Basic information: sex, birthdate, weight, and the latest heart rate and blood pressure information;

• Status of vital signs,i.e.: respiratory rate, oxygen saturation, temperature and informa- tion on physical activity and body position;

• A dynamically updated graph reﬂecting changes in selected vital signs over time;

• Triage results: those read from the analysis and inference module (initial triage) and those entered manually by the operator or medical personnel (ﬁnal triage).

The “Triage” tab presents a list of results read from the analysis and inference services. Medical Support Group Portal also allows visualization of the data using map layers providing rescuers with greater situational awareness and, thereby, facilitating planning for the prospective evacuation of victims. The symbology used as basic information indicates the triage result of each monitored soldier. In addition, operators can display a summary of the soldier’s last registered vital signs.

The “Signals” tab allows visualization of the selected biomedical signal registered by the measurement module of selected soldier. Figure [15](#bookmark49)shows an example visualization of one of the testing biomedical signal (i.e., photoplethysmogram).

The MSG portal was designed using responsive view technology that allows the interface to adapt to screen resolution and orientation. This allows efﬁcient work also with mobile devices, e.g., TRYTON’s “T5 Tactical Computer Terminal”, which is adapted to work in any terrain, under various environmental conditions.

![](/media/202412//1733390436.786739.jpeg)

**Figure 15.** MSG Portal: visualization of an example biomedical signal (soldier name: Kowalski, signal

name: Photoplethysmogram nr 1).

**8. Conclusions**

Currently, the DSS-MEDEVAC system was prepared in a form of the integrated testbed, where all the elements were implemented and tested. The tests conﬁrmed that prepared sensors measured both vital signs and biomedical signals in acceptable form, required by the analysis and inference subsystem and by medical personnel. The biomedical signals were compared with the data obtained by professional medical equipment. The rescuers as- sessed that they can be very useful for remote health monitoring of wounded soldiers and are a good supplement to the information on triage suggested by the DSS-MEDEVAC system. The system efﬁciency will next be assessed in ﬁeld tests to conﬁrm all the required capabilities.

**Author Contributions:** Conceptualization, P.L., J.K., T.S. and A.P.D.; methodology, P.L., J.K., A.P.D. and P.O.; software, A.P.D., L.A., J.K., T.S. and W.Z.; validation, P.L., J.K., T.S., A.P.D., P.O., L.A. and W.Z.; formal analysis, P.L., J.K., T.S. and A.P.D.; writing—original draft preparation, P.L., J.K., T.S., W.Z. and A.P.D.; writing—review and editing, P.L., J.K., T.S. and A.P.D.; supervision,J.K.; project administration, P.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Centre for Research and Development (NCBR)—Poland, under the project No. DOB-SZAFIR/09/B/006/01/2021.

**Institutional Review Board Statement:** Not applicable. **Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conﬂicts of Interest:** The authors declare no conﬂict of interest. The fundershad no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or

in the decision to publish the results.

**References**

1. Majumder, S.; Mondal, T.; Deen, M.J. Wearable Sensors for Remote Health Monitoring. Sensors **2017**, 17, 130. [[CrossRef]](https://doi.org/10.3390/s17010130)

2. Shelagh Dolan, the Technology, Devices, and Beneﬁts of Remote Patient Monitoring in the Healthcare Industry, Remote Patient Monitoring Trends & Health Devices in 2021. Available online: [https://www.insiderintelligence.com/insights/remote-patient-](https://www.insiderintelligence.com/insights/remote-patient-monitoring-industry-explained/) [monitoring-industry-explained/](https://www.insiderintelligence.com/insights/remote-patient-monitoring-industry-explained/) (accessed on 24 April 2023).

3. Viderman, D.; Seri, E.; Aubakirova, M.; Abdildin, Y.; Badenes, R.; Bilotta, F. Remote Monitoring of Chronic Critically Ill Patients after Hospital Discharge: A Systematic Review. J. Clin. Med. **2022**, 11, 1010. [[CrossRef]](https://doi.org/10.3390/jcm11041010)

4. Seshadri, D.R.; Li, R.T.; Voos, J.E.; Rowbottom,J.R.; Alfes, C.M.; Zorman, C.A.; Drummond, C.K. Wearable sensors for monitoring the internal and external workload of the athlete. NPJ Digit. Med. **2019**, 2, 71. [[CrossRef] [](https://doi.org/10.1038/s41746-019-0149-2)[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/31372506)

5. Chatterjee, T.; Bhattacharyya, D.; Yadav, A.; Pal, M. Quantiﬁcation of physiological and mental workloads of faster and slower ﬁnishers of a long-distance military training activity. BMJ Mil. Health **2022**, e002154. [[CrossRef] [](https://doi.org/10.1136/military-2022-002154)[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/36283744)

6. Kim,J.H.; Roberge, R.; Powell,J.B.; Shafer, A.B.; Williams, W.J. Measurement accuracy of heart rate and respiratory rate during graded exercise and sustained exercise in the heat using the Zephyr BioHarness. Int. J. Sports Med. **2013**, 34, 497–501. [[CrossRef]](https://doi.org/10.1055/s-0032-1327661) [[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/23175181)

7. Russell,M.; Sparkes,W.; Northeast,J.; Cook,C.J.; Love,T.D.; Bracken,R.M.; Kilduff,L.P. Changes in Acceleration and Deceleration

Capacity Throughout Professional Soccer Match-Play. J. Strength Cond. Res. **2016**, 30, 2839–2844. [[CrossRef] [](https://doi.org/10.1519/JSC.0000000000000805)[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/25474342)

8. Johnson, M.C.; Alarhayem, A.; Convertino, V.; Carter, R., 3rd; Chung, K.; Stewart, R.; Myers,J.; Dent, D.; Liao, L.; Cestero, R.; et al. Compensatory Reserve Index: Performance of a Novel Monitoring Technology to Identify the Bleeding Trauma Patient. Shock **2018**, 49, 295–300. [[CrossRef]](https://doi.org/10.1097/SHK.0000000000000959)

9. Kellett,J.; Holland, M.; Candel, B.G.J. Using Vital Signs to Place Acutely Ill Patients Quickly and Easily into Clinically Helpful Pathophysiologic Categories: Derivation and Validation of Eight Pathophysiologic Categories in Two Distinct Patient Populations of Acutely Ill Patients. J. Emerg. Med. **2023**, 64, 136–144. [[CrossRef]](https://doi.org/10.1016/j.jemermed.2022.12.024)

10. NATO First Responder APK, Defense Health Agency. Available online: [https://apkcombo.com/nato-ﬁrst-responder/mil.dha.](https://apkcombo.com/nato-first-responder/mil.dha.firstresponder) [ﬁrstresponder](https://apkcombo.com/nato-first-responder/mil.dha.firstresponder)(accessed on 15 May 2023).

11. OpenAhlta, Defense Health Agency. Available online: [https://www.health.mil/Military-Health-Topics/Technology/Support-](https://www.health.mil/Military-Health-Topics/Technology/Support-Areas/Open-Source-Technologies) [Areas/Open-Source-Technologies](https://www.health.mil/Military-Health-Topics/Technology/Support-Areas/Open-Source-Technologies)(accessed on 15 May 2023).

12. BATDOK, Air Force Research Laboratory. Available online: [https://techlinkcenter.org/technologies/battleﬁeld-assisted-trauma-](https://techlinkcenter.org/technologies/battlefield-assisted-trauma-distributed-observation-kit-batdok-software-tools/d6074edb-dde4-4a0b-8c40-c4e844917b28) [distributed-observation-kit-batdok-software-tools/d6074edb-dde4-4a0b-8c40-c4e844917b28](https://techlinkcenter.org/technologies/battlefield-assisted-trauma-distributed-observation-kit-batdok-software-tools/d6074edb-dde4-4a0b-8c40-c4e844917b28) (accessed on 15 May 2023).

13. START Adult Triage Algorithm. Radiation Emergency Medical Management: REMM (US Department of Health and Human Services). Available online: <http://www.remm.nlm.gov/startadult.htm> (accessed on 24 April 2023).

14. Federal Interagency Committee on EMS. National Implementation of the Model Uniform Core Criteria for Mass Casualty Incident Triage. 2014. Available online: [https://www.ems.gov/assets/National\_Implementation\_Model\_Uniform\_Core\_Criteria\_Mass\_](https://www.ems.gov/assets/National_Implementation_Model_Uniform_Core_Criteria_Mass_Casualty_Incident_Triage_Mar2014.pdf) [Casualty\_Incident\_Triage\_Mar2014.pdf](https://www.ems.gov/assets/National_Implementation_Model_Uniform_Core_Criteria_Mass_Casualty_Incident_Triage_Mar2014.pdf) (accessed on 24 April 2023).

15. Hinson,J.S.; Martinez, D.A.; Cabral, S.; George, K.; Whalen, M.; Hansoti, B.; Levin, S. Triage performance in emergency medicine:

A systematic review. Ann. Emerg. Med. **2019**, 74, 140–152. [[CrossRef]](https://doi.org/10.1016/j.annemergmed.2018.09.022)

16. Kemp, K.; Alakare, J.; Harjola, V.P.; Strandberg, T.; Tolonen,J.; Lehtonen, L.; Castr²n, M. National Early Warning Score 2 (NEWS2) and 3-level triage scale as risk predictors in frail older adults in the emergency department. BMC Emerg. Med. **2020**, 20, 83. [[CrossRef]](https://doi.org/10.1186/s12873-020-00379-y)

17. Kikomeko, B.; Mutiibwa, G.; Nabatanzi, P.; Lumala, A.; Kellett,J.A. A prospective, internal validation of an emergency patient triage tool for use in a low resource setting. Afr. J. Emerg. Med. **2022**, 12, 287–292. [[CrossRef]](https://doi.org/10.1016/j.afjem.2022.05.003)

18. Architecture Capability Team, Consultation, Command & Control Board, NATO Architecture Framework Version 4, January

2018. Available online: [https://www.nato.int/nato\_static\_ﬂ2014/assets/pdf/2021/1/pdf/NAFv4\_2020.09.pdf](https://www.nato.int/nato_static_fl2014/assets/pdf/2021/1/pdf/NAFv4_2020.09.pdf) (accessed on

24 April 2023).

19. Bazyar, J.; Farrokhi, M.; Khankeh, H. Triage Systems in Mass Casualty Incidents and Disasters: A Review Study with a Worldwide Approach. Open Access Maced. J. Med. Sci. **2019**, 7, 482–494. [[CrossRef]](https://doi.org/10.3889/oamjms.2019.119)

20. Zhao, F.; Zhao, C.; Bai, S.; Yao, L.; Zhang, Y. Triage Algorithms for Mass-Casualty Bioterrorism: A Systematic Review. Int. J. Environ. Res. Public Health **2023**, 20, 5070. [[CrossRef](https://doi.org/10.3390/ijerph20065070)] [[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/36981980)

21. Park,J.Y. Real-Time Monitoring Electronic Triage Tag System for Improving Survival Rate in Disaster-Induced Mass Casualty Incidents. Healthcare **2021**, 9, 877. [[CrossRef](https://doi.org/10.3390/healthcare9070877)] [[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/34356255)

22. Polley, C.; Jayarathna, T.; Gunawardana, U.; Naik, G.; Hamilton, T.; Andreozzi, E.; Bifulco, P.; Esposito, D.; Centracchio, J.;

Gargiulo, G. Wearable Bluetooth Triage Healthcare Monitoring System. Sensors **2021**, 21, 7586. [[CrossRef] [](https://doi.org/10.3390/s21227586)[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/34833659)

23. Harold, R. Montgomery, Tactical Combat Casualty Care, Quick Reference Guide. 2017. Available online: [https://emsa.ca.gov/](https://emsa.ca.gov/wp-content/uploads/sites/71/2017/07/TCCC_Quick_Reference_Guide_2017.pdf) [wp-content/uploads/sites/71/2017/07/TCCC\_Quick\_Reference\_Guide\_2017.pdf](https://emsa.ca.gov/wp-content/uploads/sites/71/2017/07/TCCC_Quick_Reference_Guide_2017.pdf)(accessed on 24 April 2023).

24. Athaya, T.; Choi, S. A Review of Noninvasive Methodologies to Estimate the Blood Pressure Waveform. Sensors **2022**, 22, 3953. [[CrossRef](https://doi.org/10.3390/s22103953)] [[PubMed]](https://www.ncbi.nlm.nih.gov/pubmed/35632360)

25. Mukkamala, R.; Stergiou, G.S.; Avolio, A.P. Cufﬂess Blood Pressure Measurement. Annu. Rev. Biomed. Eng. **2022**, 24, 203–230. [[CrossRef]](https://doi.org/10.1146/annurev-bioeng-110220-014644)

26. Zhou, Z.B.; Cui, T.R.; Li, D.; Jian,J.M.; Li, Z.; Ji, S.R.; Li, X.; Xu,J.D.; Liu, H.F.; Yang, Y.; et al. Wearable Continuous Blood Pressure Monitoring Devices Based on Pulse Wave Transit Time and Pulse Arrival Time: A Review. Materials **2023**, 16, 2133. [[CrossRef]](https://doi.org/10.3390/ma16062133)

27. Sondej, T.; Sieczkowski, K.; Olszewski, R.; Dobrowolski, A. Simultaneous multi-site measurement system for the assessment of pulse wave delays. Biocybern. Biomed. Eng. **2019**, 39, 488–502. [[CrossRef]](https://doi.org/10.1016/j.bbe.2019.01.001)

28. Plesinger, F.; Jurco, J.; Halamek,J.; Jurak, P. SignalPlant: An open signal processing software platform. Physiol. Meas. **2016**, 37, N38–N48. [[CrossRef]](https://doi.org/10.1088/0967-3334/37/7/N38)

29. Multidomain HMS C3IS JASMINE. Available online: <https://defence24.com/industry/multidomain-hms-c3is-jasmine>(accessed on 24 April 2023).

30. Google Remote Procedure Calls—A High Performance, Open Source Universal RPC Framework. Available online: [https:](https://grpc.io/) [//grpc.io/](https://grpc.io/) (accessed on 24 April 2023).

31. Pollaris, G.; Sabbe, M. Reverse triage: More than just another method. Eur. J. Emerg. Med. **2016**, 23, 240–247. [[CrossRef]](https://doi.org/10.1097/MEJ.0000000000000339)

32. W6jtowicz, B.; Dobrowolski, A.P.; Tomczykiewicz, K. Fall detector using discrete wavelet decomposition and SVM classiﬁer. Metrol. Meas. Syst. **2015**, 22, 303–314. [[CrossRef]](https://doi.org/10.1515/mms-2015-0026)

33. Dobrowolski, A.P.; Suchocki, M.; Tomczykiewicz, K.; Majda-Zdancewicz, E. Classiﬁcation of auditory brainstem response using wavelet decomposition and SVM network. Biocybern. Biomed. Eng. **2016**, 36, 427–436. [[CrossRef]](https://doi.org/10.1016/j.bbe.2016.01.003)

34. Stasiakiewicz, P.; Dobrowolski, A.P.; Targowski, T.; Gała˛zka- widerek, N.; Sadura-Sieklucka, T.; Majka, K.; Skoczylas, A.;

Lejkowski, W.; Olszewski, R. Automatic classiﬁcation of normal and sick patients with crackles using wavelet packet decomposi- tion and support vector machine. Biomed. Signal Process. Control **2021**, 67, 102521. [[CrossRef]](https://doi.org/10.1016/j.bspc.2021.102521)

**Disclaimer/Publisher’s Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.