"An Emotional Support Animal, Without the Animal": Design Guidelines for a Social Robot to Address Symptoms of Depression
"没有动物的情感支持动物"：解决抑郁症症状的社交机器人设计指南

ABSTRACT
摘要

Socially assistive robots can be used as therapeutic technologies to address depression symptoms. Through three sets of workshops with individuals living with depression and clinicians, we developed design guidelines for a personalized therapeutic robot for adults living with depression. Building on the design of Therabot , workshop participants discussed various aspects of the robot’s design, sensors, behaviors, and a robot connected mobile phone app. Similarities among participants and workshops included a preference for a soft textured exterior and natural colors and sounds. There were also differences - clinicians wanted the robot to be able to call for aid, while participants with depression differed in their degree of comfort in sharing data collected by the robot with clinicians.
社交辅助机器人可用作治疗技术，以解决抑郁症状。通过与抑郁症患者和临床医生举行三组研讨会，我们制定了针对成年抑郁症患者的个性化治疗机器人的设计指南。以 Therabot 的设计为基础，研讨会的参与者讨论了机器人的设计、传感器、行为以及与机器人连接的手机应用等各个方面。与会者和研讨会的相同之处包括：他们更喜欢质地柔软的外观、自然的颜色和声音。但也有不同之处--临床医生希望机器人能够呼救，而患有抑郁症的与会者则在与临床医生分享机器人收集的数据时表现出不同的舒适度。

CCS CONCEPTS
综合传播战略概念

Human-centered computing HCI design and evaluation methods; - Social and professional topics User characteristics.
以人为本的计算 人机交互设计和评估方法； - 社会和专业主题 用户特征。

KEYWORDS
关键词

Socially Assistive Robots, Depression, Human Robot Interaction, Robot Design
社交辅助机器人、抑郁症、人机交互、机器人设计

1 INTRODUCTION
1 引言

Depression is widespread, with approximately of the global population experiencing depression at some point in their lives [46]. Socially assistive robots (SARs) have shown promising therapeutic benefits within the context of depression, including a reduction in depression symptoms [3] and increased happiness [24]. Currently used robots, however, are commonly not personalized to the needs of individual users. People living with depression can have a wide array of personalities and preferences, and depression itself can involve various co-morbidities that make generalized design difficult. These co-morbidities can include other mental health disorders [40] or physical disabilities [8].
抑郁症很普遍，全球约有 的人口在一生中的某个阶段经历过抑郁症[46]。社交辅助机器人（SAR）对抑郁症有很好的治疗效果，包括减少抑郁症状[3]和增加幸福感[24]。然而，目前使用的机器人通常不能满足用户的个性化需求。抑郁症患者的个性和偏好多种多样，而且抑郁症本身可能涉及各种并发症，因此很难进行通用化设计。这些并发症可能包括其他精神疾病[40]或身体残疾[8]。

Co-design with adults with depression provides one way to become aware of and address personal differences in the design of new technologies for Human Robot Interaction (HRI). Researchers acknowledge that, when designing for a specialised population, involving members of that population directly in the design process can bring forward ideas and functions that may be beyond the expertise of researchers alone. This perspective views the user population as experts in their own experience [33]. However, when working with participants who live with physical or mental illness, special challenges arise - such as ensuring that the participant’s well-being is forefront in the research and maintaining an inclusive recruitment process.
在设计人机交互（HRI）新技术时，与患有抑郁症的成年人共同设计是了解和解决个人差异的一种方法。研究人员承认，在为特殊人群进行设计时，让该人群的成员直接参与设计过程，可以提出研究人员无法独立完成的想法和功能。这种观点认为，用户群体是其自身经验的专家[33]。然而，当与患有身体或精神疾病的参与者一起工作时，就会出现一些特殊的挑战--例如，如何确保参与者的福祉在研究中处于最重要的位置，以及如何保持招募过程的包容性。

To better understand the potential and need for personalized design of the socially assistive robot (SAR) Therabot (see figure 1) as a companion to traditional therapy for those living with depression, we developed three related workshop series. Discussions and activities in the workshops focused on the appearance of the robot, the behaviors and sounds of the robot, the sensors onboard the robot, any privacy concerns regarding the data collected by the robot, and the usefulness of a robot connected phone app. Five participants recruited from the mental health provider Centerstone, who had a confirmed diagnosis of depression, took part in a five-part online workshop covering the above topics. While recruitment with the facility Centerstone took place, we also recruited those from outside the facility. Ten participants who self-reported a depression diagnosis took part in a one-hour-long single online workshop. By recruiting in both ways, this allowed participants who may not have an official depression diagnosis from Centerstone to participate, as well as allowed us to test a shortened version of the workshops. Finally, four clinicians participated in four online workshops in which they suggested design guidelines for SARs for those living with depression.
为了更好地了解社交辅助机器人（SAR）Therabot （见图 1）作为抑郁症患者传统疗法的辅助工具的潜力和个性化设计需求，我们举办了三个相关的系列研讨会。工作坊中的讨论和活动主要集中在机器人的外观、机器人的行为和声音、机器人上的传感器、机器人收集数据的隐私问题以及机器人连接手机应用的实用性。从心理医疗机构 Centerstone 招募的五名确诊为抑郁症的参与者参加了由五个部分组成的在线研讨会，内容涵盖上述主题。在中心石机构进行招募的同时，我们还从机构外招募了参与者。十名自我报告确诊为抑郁症的参与者参加了一个小时的单次在线研讨会。通过这两种方式进行招募，可以让那些可能没有从中心石获得正式抑郁症诊断的参与者参与进来，同时也可以让我们对缩短版的工作坊进行测试。最后，四名临床医生参加了四次在线研讨会，并在研讨会上提出了针对抑郁症患者的 SAR 设计指南。

By working with these diverse groups of participants, we aimed to find common themes regarding the robot’s design to be implemented as options for users living with depression who may receive the Therabot robot in their homes.
通过与这些不同的参与者群体合作，我们旨在找到有关机器人设计的共同主题，并将其作为可能在家中使用 Therabot 机器人的抑郁症患者的选择。

Through the robot Therabot we sought to explore the effects of technology as similar to the calming nature of animal assisted therapy [?], and the ability of data collection through sensors for management depression symptoms [12]. We present those themes below, including areas of the physical design of the robot, use cases for the robot, and concerns regarding data collection. Using these themes, we further discuss how participants requests may be implemented into the robot itself.
通过机器人Therabot 我们试图探索技术的效果是否类似于动物辅助治疗的镇静作用[?]，以及通过传感器收集数据来管理抑郁症状的能力[12]。我们将在下文中介绍这些主题，包括机器人的物理设计、机器人的使用案例以及数据收集方面的问题。利用这些主题，我们将进一步讨论如何将参与者的要求落实到机器人本身。

Figure 1: Therabot with and without covering
图 1：有覆盖物和无覆盖物的 Therabot

2 RELATED WORK
2 相关工作

In the following sections, we review prior work on depression and its symptoms and common treatments, the use of SARs in mental health contexts generally and for depression specifically, and the participatory design of SARs.
在下面的章节中，我们将回顾之前关于抑郁症及其症状和常见治疗方法的工作、在精神健康环境中使用简化报告的一般情况和抑郁症的具体情况，以及简化报告的参与式设计。

2.1 Depression Prevalence, Symptoms and Tools
2.1 抑郁症患病率、症状和工具

Depression is one of the leading mental illnesses in the world with approximately 280 million people living with the condition [46]. In 2021 alone, an estimated 21 million adults over the age of 18 had at least one depressive episode within the United States [35]. Depression can involve a variety of symptoms that cause both mental and physical strain on individuals, such as depressed mood, feeling worthless, suicidal ideation, and fatigue [18]. In addition to the symptoms commonly associated with depression, those who experience the illness may also experience co-morbidities brought on or exacerbated by their condition. This can include other diagnosed mental illnesses, such as generalized anxiety disorder [40] or physical conditions such as heart disease [8], which can produce difficulties in treating the individual holistically. There are many forms of treatment for depression, including working with a therapist, a psychiatrist, or a general health provider such as a primary care physician. One common therapy includes cognitive behavioral therapy which encourages cognitive restructuring and belief change [43]. Virtual sessions with a therapist have become more prominent, with an increase in individuals seeking this care [27], demonstrating its potential to make therapy more easily accessible [26]. Other therapies exist outside of a typical therapist’s office or clinic, including animal assisted therapy [21], in which a non-human animal works with the individual. Animal-assisted therapy has been shown to reduce feelings of depression [25], through the encouragement of the participant to talk to the animal, and engaging in play with the animal [7].
抑郁症是全球主要精神疾病之一，约有 2.8 亿人患有抑郁症[46]。仅在 2021 年，估计美国就有 2,100 万 18 岁以上的成年人至少有一次抑郁发作 [35]。抑郁症会出现多种症状，给患者造成精神和身体上的双重压力，如情绪低落、感觉自己毫无价值、有自杀倾向和疲劳[18]。除了抑郁症常见的相关症状外，抑郁症患者还可能出现因其病情而引发或加重的并发症。这可能包括其他已确诊的精神疾病，如广泛性焦虑症[40]或躯体疾病，如心脏病[8]，这些都会给个体的整体治疗带来困难。抑郁症的治疗有多种形式，包括与治疗师、精神科医生或全科医生（如初级保健医生）合作。一种常见的疗法包括认知行为疗法，这种疗法鼓励认知重组和信念改变 [43]。与治疗师的虚拟会话已变得越来越重要，越来越多的人寻求这种治疗[27]，这表明虚拟会话有可能使治疗更容易获得[26]。在典型的治疗师办公室或诊所之外还有其他疗法，包括动物辅助疗法[21]，即由非人类动物与患者一起工作。动物辅助疗法通过鼓励参与者与动物交谈，并与动物一起玩耍，已被证明可以减轻抑郁情绪[25]。

2.2 Socially Assistive Robots
2.2 社会辅助机器人

SARs function as robots that provide social interaction to support the user [14]. These robots have been used for a variety of applications, including motivation (encouraging the user’s behaviors) [1], coaching (leading sessions) [19], and companionship [13]. The humanoid NAO robot has been used to alleviate anxiety with children in a hospitalization setting, showing a decrease in cortisol levels [39]. The NAO robot has also been used to develop and encourage communication skills with children diagnosed with autism [42].
合成孔径雷达是一种提供社交互动以支持用户的机器人[14]。这些机器人已被用于多种用途，包括激励（鼓励用户的行为）[1]、辅导（引导课程）[19]和陪伴[13]。人形机器人NAO曾被用于缓解住院儿童的焦虑情绪，结果显示皮质醇水平有所下降[39]。NAO 机器人还被用于培养和鼓励被诊断患有自闭症的儿童的交流技能 [42]。

One of the most ubiquitous SARs is the robotic seal Paro, which has been used with older adults living with dementia [44]. Paro’s use led to increasing interaction between the user and their care team [41], as well as a reduction in depression and agitation [22].
机器人海豹 Paro 是最常见的 SAR 之一，它已被用于患有痴呆症的老年人[44]。帕罗的使用增加了用户与其护理团队之间的互动[41]，并减少了抑郁和躁动[22]。

As the number of uses cases for SARs grows, engaging each unique target population in the design process is important for increasing inclusiveness in robot design. Participatory design and co-design, provide a method where users are directly involved in the design process [30]. One such area is the development of a SAR to be used in education, where aspects such as communication features and appearance were explored by stakeholders [37].
随着合成孔径雷达使用案例的增加，让每个独特的目标人群参与设计过程对于提高机器人设计的包容性非常重要。参与式设计和共同设计提供了一种让用户直接参与设计过程的方法[30]。其中一个例子是开发用于教育领域的 SAR，相关人员对其交流功能和外观等方面进行了探讨 [37]。

Particular to depression, researchers have also completed participatory design workshops with older adults living with depression which explored the overall look of the SAR as well as potential uses [48]. These studies implemented various forms of data collection, from writing and discussion to sketching.
针对抑郁症，研究人员还与患有抑郁症的老年人共同完成了参与式设计工作坊，探讨了特区的整体外观和潜在用途[48]。这些研究采用了各种形式的数据收集，从写作、讨论到素描。

2.2.1 SARs used for Depression management. Many studies have focused on the use of socially assistive robots and their effects on depression with older adults specifically. In one study, older adults living with depression received Paro to use in their homes, while being monitored by researchers [38]. It was found that those that used Paro had a reduction in depression symptoms and the sensors placed on the robot aided in predicting depression levels [3].
2.2.1用于抑郁管理的 SAR。许多研究关注社交辅助机器人的使用及其对老年人抑郁症的具体影响。在一项研究中，患有抑郁症的老年人可以在家中使用 Paro，同时接受研究人员的监控 [38]。研究发现，使用帕罗的老年人抑郁症状有所减轻，机器人上的传感器有助于预测抑郁程度[3]。

Within the context of a care home in Taiwan the robot Paro also showed reduced depression and an improvement of quality of life, with participants humanizing the robot and using it as a way to engage with others [9]. Many studies focused on the SAR Paro, though less than half reported by Araujo et al. found improvement of depression symptoms [2]. Other studies, with more positive outcomes, included older adults with depression but developed a unique robot outside of what is commercially available, such as eBear [24] and found that the robot could increase the happiness of and uplift the individuals interacting with it. While one research group created Ryan, a robot for delivering cognitive behavioral therapy to older adults as an alternative to human only interaction to positive reviews [15].
在台湾的一家护理院中，机器人帕罗也显示抑郁症有所减轻，生活质量有所提高，参与者将机器人人性化，并将其作为与他人接触的一种方式[9]。许多研究都集中在特区帕罗上，但 Araujo 等人的报告显示，只有不到一半的研究发现抑郁症状有所改善[2]。其他研究则取得了更积极的成果，研究对象包括患有抑郁症的老年人，但开发出了一种独特的机器人，而非市面上销售的机器人，如 eBear [24]，研究发现这种机器人可以提高与之互动的人的幸福感，并使他们振奋起来。有一个研究小组创造了 Ryan 机器人，为老年人提供认知行为疗法，作为人机互动的替代方案，获得了积极的评价[15]。

3 CO-DESIGN WORKSHOPS
3 次共同设计研讨会

Three workshop types were conducted as part of this study, one with recruitment through Centerstone, one with those who self reported receiving a depression diagnosis elsewhere, and one with clinicians. Participants who were diagnosed with depression were recruited through flyers at the mental health facilities Centerstone and word of mouth from clinicians. Clinicians were recruited exclusively through word of mouth by office personnel who were familiar with the study. For the short form workshops, participants were recruited through Indiana University Classifieds system. Each of these workshops were approved by the Indiana University Institutional Review Board.
作为本研究的一部分，我们举办了三种类型的研讨会，一种是通过中心石招募，一种是自我报告在其他地方接受抑郁症诊断的人参加，还有一种是临床医生参加。被诊断为抑郁症的参与者是通过中心石心理健康机构的传单和临床医生的口碑招募的。临床医生完全是通过熟悉研究的办公室人员的口口相传招募的。简短研讨会的参与者则是通过印第安纳大学分类系统招募的。每场研讨会都获得了印第安纳大学机构审查委员会的批准。

3.1 Long Form Workshops
3.1 长篇研讨会

Participants recruited through Centerstone for the long form workshops were invited to take part in five total virtual group meetings, each lasting an hour. These workshops were performed online through Zoom. Before the workshops began, participants were invited to a brief meeting with one of the researchers to discuss the study, ask any questions they may have, and to collect verbal consent, demographic data, and responses on the PhQ-9, a depression assessment questionnaire [28], from participants. The PhQ-9 rates the intensity of depression symptoms over the previous two weeks, with 5-9 being mild, 10 - 14 being moderate, 15-19 is moderately severe and over 20 is severe.
通过中心石招募的长篇研讨会参与者受邀参加了总共五次虚拟小组会议，每次会议持续一小时。这些研讨会通过 Zoom 在线进行。在工作坊开始之前，参与者会被邀请与其中一位研究人员进行简短的会面，讨论研究内容，询问他们可能有的任何问题，并收集参与者的口头同意、人口统计学数据以及对抑郁评估问卷 PhQ-9 的回答[28]。PhQ-9对过去两周的抑郁症状强度进行评分，5-9分为轻度，10-14分为中度，15-19分为中度重度，20分以上为重度。

Workshop 1 provided an introduction to socially assistive robots and presented participants with two commercial robots through video (Buddy [6], and Paro[36]). They were also introduced to the robot Therabot that they would be focusing on during the following workshops, as well as basic sensors that may be present in the robot (such as ambient light, sound, and touch).
工作坊 1 介绍了社交辅助机器人，并通过视频向与会者展示了两款商用机器人（Buddy [6] 和 Paro[36]）。他们还了解了机器人Therabot （他们将在接下来的工作坊中重点关注），以及机器人中可能存在的基本传感器（如环境光、声音和触摸）。

In workshop 2 a sketch artist from Mississippi State University joined. During this workshop sensors were reviewed, and participants were asked about their preference on the physical design and outer covering of the four-legged Therabot design. While participants described their desired robot covering, the sketch artist drew alongside, before showing the participants the drawing which was then altered based on participant specifications.
在研讨会 2 中，来自密西西比州立大学的素描艺术家加入了研讨会。在这次工作坊中，对传感器进行了检查，并询问了参与者对四条腿的 Therabot 设计的物理设计和外部覆盖物的偏好。当参与者描述他们想要的机器人外罩时，素描艺术家在旁边作画，然后向参与者展示图纸，再根据参与者的要求对图纸进行修改。

In workshop 3, participants were asked to sketch the layout of their home environment, how they move through their space, and where they spend the most time. They were then asked about where they thought they might use the robot the most, and why it may be used within those specific areas. Workshop 4 turned focus from the physical robot to a phone application that may be paired with the robot. Participants were asked to describe and draw what they would want to be able to do with the application, both when the robot is present and when it is not.
在研讨会 3 上，与会者被要求勾画出他们的家居环境布局、他们如何在空间中移动以及他们在哪里花费的时间最多。然后询问他们认为在哪些地方最需要使用机器人，以及在这些特定区域使用机器人的原因。研讨会 4 将重点从实体机器人转向可能与机器人配对的手机应用程序。与会者被要求描述并绘制他们希望在机器人在场或不在场的情况下使用该应用程序所能做的事情。

Workshop 5 was a final review, where participants were asked to read a brief story of an individual who lives with depression, and then re-imagine what that individual’s daily life may look like with the robot that they had designed, based upon the previous 4 workshops. The prompts included one where the individual was at home, and one where the individual was in a therapy setting (such as a therapist’s office). Participants were asked to write a short story using the prompt that focused on the individual in the prompt using their SAR.
工作坊 5 是最后的回顾，要求参与者阅读一个患有抑郁症的人的简短故事，然后根据前面 4 个工作坊的内容，重新想象这个人与他们设计的机器人在一起时的日常生活。提示包括一个人在家中的提示和一个人在治疗环境（如治疗师办公室）中的提示。参与者被要求根据提示写一个小故事，故事的重点是提示中的个人使用他们的特区。

In all five workshops, participants were invited to volunteer any information they thought about between sessions, and were asked to consider privacy and data collection implications when using the robot or phone application. At the end of each workshop participants were remunerated for their work.
在所有五场研讨会上，与会者都被邀请在会议间隙自愿提供他们想到的任何信息，并被要求在使用机器人或手机应用程序时考虑隐私和数据收集的影响。每次研讨会结束后，与会者都会获得相应的报酬。

3.2 Short Form Workshops
3.2 短期讲习班

The short form workshops covered, briefly, much the same information as the long form workshops, as reported in [11]. This workshop took place over a period of an hour long video session, through Zoom. Participants were given a demographic questionnaire as well as the PhQ-9. Then, they were first introduced to SARs through Joy for All and Paro through videos, then they were introduced to Therabot via picture.
如文献[11]所述，短期研讨会简要介绍了与长期研讨会大致相同的信息。该研讨会通过 Zoom 视频会议进行，时长一小时。参与者会收到一份人口统计学问卷以及 PhQ-9。然后，首先通过视频向他们介绍了 "全民悦动 "和 "Paro"，接着又通过图片向他们介绍了 Therabot。

Focusing on the outside covering of Therabot, participants were first asked to describe the covering of the robot based upon the skeletal structure. They were asked to draw two coverings based upon their preference, with one being focused on someone who did not like cats or dogs. Once they had chosen a covering to focus on, they were then introduced to common sensors, and asked what sensors they would prefer the robot include. Next participants were asked about the sounds and behaviors that they would want from the robot that they had designed, with specific questions focused on therapeutic tools.
针对 Therabot 的外部覆盖物，首先要求参与者根据骨骼结构描述机器人的覆盖物。他们还被要求根据自己的喜好画出两种外罩，其中一种外罩是针对不喜欢猫或狗的人的。一旦他们选择了一种覆盖物，就会向他们介绍常见的传感器，并询问他们希望机器人包含哪些传感器。接下来，参与者被问及他们希望自己设计的机器人发出什么样的声音和做出什么样的行为，具体问题主要集中在治疗工具上。

Finally participants were asked about a robot connected phone app, and what capabilities they would want the app to have as well as what information they would want present. Particularly related to sensors and tracking information, participants were also asked about any privacy concerns that they may have when using the robot or the app. Last, participants were asked about any concerns they may have with using the robot.
最后，与会者被问及与机器人连接的手机应用程序，以及他们希望该应用程序具备哪些功能和提供哪些信息。特别是在传感器和跟踪信息方面，参与者还被问及在使用机器人或应用程序时可能会有的隐私问题。最后，与会者还被问及他们在使用机器人时可能会有的任何顾虑。

3.3 Clinician Workshops
3.3 临床医生讲习班

We performed four online clinician workshops in total, with the first three being attended by all clinicians and a fourth being attended only by clinicians that may spend time located in the clients homes.
我们共举办了四次临床医生在线研讨会，前三次研讨会由所有临床医生参加，第四次研讨会仅由可能在客户家中工作的临床医生参加。

Workshop 1 introduced the clinicians to socially assistive robots via description and video examples (Buddy[6], Mabu[31], Paro[36]). The clinicians were encouraged to discuss what they did and did not like about the robots for their clients, before being introduced to Therabot . Next the clinicians were asked to play a game of "Yes, Let’s..." an improvisation game where they each indicated a single aspect they wished to add to a socially assistive robot, by going around in a circle and building on the previous participants answer. This exercise was repeated twice: first, to discuss the physical design of the robot and second, to address the features they wanted the robot to sense, following a brief introduction to basic sensors.
工作坊 1 通过描述和视频示例（Buddy[6]、Mabu[31]、Paro[36]）向临床医生介绍了社交辅助机器人。在介绍 Therabot 之前，我们鼓励临床医生讨论他们喜欢和不喜欢客户使用的机器人。接下来，临床医生们被要求玩一个 "是的，让我们...... "的即兴游戏。在这个游戏中，他们围成一圈，在前一位参与者回答的基础上，各自说出希望为社交辅助机器人添加的一个方面。这一练习重复了两次：第一次是讨论机器人的物理设计，第二次是在简单介绍基本传感器之后，讨论他们希望机器人能够感知的功能。

In workshop 2 clinicians were introduced to the robot Therabot and its current touch sensors via video. Then they were shown robot designs from three clients that had completed the long form workshops. They were asked to give feedback on the clients’ designs, as well as add what they would like to them. They were then asked questions regarding privacy and data collection with the robot, particularly in relation to sensors that they had chosen in the previous workshop as well as more specific sensors like a camera.
在工作坊 2 中，临床医生通过视频了解了机器人 Therabot 及其当前的触摸传感器。然后，向他们展示了三位完成了长期工作坊的客户的机器人设计。他们被要求对客户的设计提出反馈意见，并添加他们想要的内容。然后，他们被问到了有关隐私和机器人数据收集的问题，特别是与他们在之前的工作坊中选择的传感器以及更具体的传感器（如摄像头）相关的问题。

Table 1: Centerstone Diagnosed Demographic Data.
表 1：Centerstone 诊断人口数据。

In workshop 3, clinicians focused on the robot connected phone app. They were asked about their initial thoughts on the app, as well as shown a few designs from two of the client participants to discuss some of the requests from the the long form workshops. Finally, in workshop 4 only the clinicians that have spent time traveling to clients homes were included. They briefly discussed how they thought the robot might fit into the home environment of some of their clients with depression.
在第三场研讨会上，临床医生重点讨论了与机器人连接的手机应用。他们被问及对该应用程序的初步想法，并展示了两位客户参与者的一些设计，以讨论长期研讨会上提出的一些要求。最后，在第四场研讨会中，只有那些花时间前往客户家中的临床医生参与了讨论。他们简要讨论了如何将机器人融入到抑郁症患者的家庭环境中。

4 RESULTS
4 结果

All excerpts were inductively coded by the first author, after being auto-transcribed and manually corrected from the workshops videos by a secondary member of the research team. Within the client coding, there was an agreement in code application between the first author and a second research assistant who was familiar with the data. In this process the first author coded the workshops in their entirety, and then the secondary member of the research team reviewed each transcription, and excerpt for accuracy while coding. The that were not in agreement were discussed between the two researchers and re-coded until codes were agreed upon. These codes were then compared to the codebook from the short form workshops to find overlaps and differences in the themes spoken about by participants, which was also confirmed by a secondary researcher. Codes which had a direct link (such as the cat form factor) were combined. Within the coding of the clinician based workshop there was a agreement, with the re-coded until codes were agreed upon. See appendix for codes application and explanation for each of the workshop types.
所有摘录均由第一作者进行归纳编码，然后由研究小组的第二名成员对研讨会视频进行自动转录和人工校正。在客户编码过程中， 第一作者与熟悉数据的第二位研究助理在代码应用方面达成了一致。在此过程中，第一作者对研讨会的全部内容进行编码，然后由研究小组的第二名成员审查每份记录和摘录，以确保编码的准确性。 两位研究人员对不一致的内容进行讨论，并重新编码，直到达成一致。然后，将这些编码与简表研讨会的编码本进行比较，以发现参与者所谈主题的重叠和不同之处，这一点也得到了辅助研究人员的确认。有直接联系的代码（如猫的形式因素）被合并在一起。在以临床医生为基础的工作坊的编码中， 达成了一致， 重新编码，直到编码达成一致。每种研讨会类型的代码应用和解释见附录。

4.1 Participants
4.1 参与者

In the long form workshops, seven participants were recruited through Centerstone to participate. Of those seven, only five completed all five workshops, as one participant stopped attending due to scheduling constraints, and one stopped attending due to a mental health crisis (see Table 1). These five participants had a confirmed diagnosis of major depressive disorder from Centerstone clinicians, though many also had other co-morbid diagnoses. Participants 1 and 2 participated in the workshops together, participant five participated alone after the second workshop, and participant 6 and 7 participated together except for the final workshop due to scheduling constraints.
在长篇研讨会中，通过 Centerstone 招募了七名参与者。在这七名参与者中，只有五人完成了全部五次工作坊，其中一名参与者因日程安排的限制而停止参加，另一名参与者因心理健康危机而停止参加（见表 1）。这五名参与者均被 Centerstone 临床医生确诊为重度抑郁障碍，但其中许多人还合并有其他疾病。参与者 1 和 2 一起参加了工作坊，参与者 5 在第二次工作坊后单独参加了工作坊，参与者 6 和 7 除最后一次工作坊外因时间安排的限制一起参加了工作坊。

In the short form workshop, 10 participants took part in a 1 hour long workshop. These participants self-reported having a diagnosis of depression currently or in the past. Eight of the 10 completed the entire workshop, and two did not complete the phone app related portion (Table 2).
在简易工作坊中，10 名参与者参加了一个小时的工作坊。这些参与者自称目前或过去曾被诊断患有抑郁症。10 人中有 8 人完成了整个研讨会，2 人未完成手机应用相关部分（表 2）。

Table 2: Short Form Participant Demographic Data.
表 2：简表参与者人口统计数据。

The four clinician participants were asked to describe their role at Centerstone. This group included those that worked in outpatient recovery, care coordination, dialectic behavior focused therapies, and a team leader.
四名临床医生被要求描述他们在 Centerstone 的角色。其中包括门诊康复、护理协调、辩证行为疗法以及团队领导者。

4.2 Methodology Considerations and Comparison
4.2 方法考虑与比较

Access to therapy continues to change, both due to an increase of mental health related symptoms occurring during the COVID- 19 pandemic [20] and increased workload to clinicians [16]. By inviting those who have a previous depression diagnosis outside of the local mental health facility, we were able to provide greater access for this community to participate in the design of future technologies.
由于 COVID-19 大流行期间精神健康相关症状的增加[20]以及临床医生工作量的增加[16]，获得治疗的途径仍在不断变化。通过邀请那些曾在当地心理健康机构之外诊断出抑郁症的人，我们能够为这一群体提供更多参与未来技术设计的机会。

Despite the different participants, the core concepts within the workshops stayed the same (i.e. phsyical design of the robot, robot behaviors). However, within the short form workshops, an additional question regarding any concerns they may have regarding the use of SARs for depression was added. Within the long form workshops more time was spent going over the specific areas of the home that the robot might be used in, rather than talking about its use more generally.
尽管参与者不同，但工作坊的核心概念保持不变（即机器人的物理设计、机器人行为）。不过，在简式工作坊中，还增加了一个问题，即他们对使用合成孔径雷达治疗抑郁症是否有任何顾虑。在长式研讨会中，更多的时间用于讨论机器人可能在家中使用的具体区域，而不是更广泛地讨论机器人的使用。

Time was also an important factor in the workshops - whereas the long form workshops took place over five one-hour-long workshops where participants could focus on a singular aspect for the entire hour, in the short form workshop only one hour was allocated to cover all topics. Out of the 10 participants in the short form workshop, the time constraint prevented two participants from reporting their preferences for the robot connected app.
时间也是工作坊的一个重要因素--长式工作坊分为五个一小时的工作坊，参与者可以在整整一个小时的时间里专注于一个方面，而短式工作坊只有一个小时的时间来涵盖所有主题。在简短研讨会的 10 名参与者中，由于时间限制，有两名参与者无法报告他们对机器人连接应用程序的偏好。

Overall the long form (five workshop) and short form (single workshop) workshops provided similar main themes and requests from participants for designing a socially assistive robot based on the Therabot form.
总体而言，长式（五次研讨会）和短式（一次研讨会）研讨会提供了类似的主题，以及与会者对基于 Therabot 形式设计社交辅助机器人的要求。

In the following section the outcomes of these workshops are described as the collective responses of all participants, and the similarities and differences between the three participant groups.
下一节将介绍这些研讨会的成果，即所有与会者的集体反应，以及三个与会者群体之间的异同。

4.3 Workshops with Adults with Depression
4.3 为患有抑郁症的成年人举办讲习班

When combining the data between both the long form and short form workshops, various design guidelines start to emerge for designing a SAR to be used in combination with other therapeutic methods. Below," " refers to all 15 participants who took part in either the long (confirmed diagnosis of depression) or short form (self-reported diagnosis of depression) workshops.
style="font-family: "Times New Roman"; min-height: 11pt; font-size: 11pt;" data-imersive-translate-walked="287aa928-15d3-4c1f-b165-c43cec8e695d">当把长表和短表工作坊的数据结合起来时，开始出现各种设计准则，用于设计与其他治疗方法结合使用的特别研究区。以下，"n "指所有参加了长式（确诊为抑郁症）或短式（自我报告诊断为抑郁症）工作坊的 15 位参与者。

4.3.1 Robot Physical Design and Sensors. When discussing the robot’s appearance, after previously having been shown the Therabot robot in its beagle form or uncovered, participants decided that a more familiar form would be preferable. The most popular was a cat , with the second overall most popular being a dog ), with other variations (such as an bear) only being requested by one participant (figure 1). This could be due to the participants’ previous relationships with cats and dogs.
4.3.1 机器人物理设计和传感器。在讨论机器人的外观时，在之前展示过 Therabot 机器人的小猎犬形态或未遮盖的形态之后，参与者认为更熟悉的形态会更好。最受欢迎的是猫 ，其次是狗 ），只有一名参与者要求其他变体（如熊）（图 1）。这可能与参与者以前与猫和狗的关系有关。

For example, one long form workshop participant described a cat she had as a child as inspiration for her robot design (LFP1): "Very fluffy the fluff. When I was growing up as a kid, I had a cat named Lovey who was just a big fluff ball. And you know, I had a lot of emotional problems as a kid and I would like sometimes just pick her up and like her whole fluff would be against me and sometimes I would cry into her:"
例如，一位长篇研讨会与会者描述了她小时候养的一只猫，这是她设计机器人的灵感来源（LFP1）："非常蓬松的绒毛。我小时候养了一只叫 Lovey 的猫，它就是一个大绒球。你知道，我小时候有很多情绪问题，有时候我把她抱起来，她的绒毛就会紧紧贴着我，有时候我会对着她大哭一场："

Figure 2: Sketch Artist Designs
图 2：素描艺术家的设计

Relatedly, natural colors were the most requested , with some participants referencing the colors of animals that they were familiar with. LFP7 explained, "I have another idea that I think I would be more interested in seeing a sketch of. OK, you know what a tortoise shell cat looks like? I love torties."
与此相关的是，自然色彩是最受欢迎的 ，一些参与者提到了他们熟悉的动物的颜色。LFP7 解释说："我有另一个想法，我想我更想看看草图。你知道龟壳猫长什么样吗？我喜欢龟壳猫。

Participants also indicated that the texture of the robot was important to them, while they had different ideas on the material (such as specific types of fur) many used the term "soft" when describing their desired texture . LFP7 said,"I think that plush texture would be really important. OK. So for me, it’s important to have it, you know, it’s soft. The user wants to touch it." SFP8 described, "I would want it to be soft and cuddly like a Build a Bear."
与会者还表示，机器人的质地对他们很重要，虽然他们对材料（例如特定类型的毛皮）有不同的想法，但许多人在描述他们想要的质地时使用了 "柔软 "一词 。LFP7 说："我认为毛绒质感非常重要。好的。对我来说，最重要的是它要柔软。用户想要触摸它"。SFP8描述道："我希望它像Build a Bear一样柔软可爱。

The focus on a soft and touchable covering provides context for participants’ most requested sensing capability, which was the robot being able to process touch . LFP6 explained,"Having something to... talk to and you know cuddle with and touch, that would be nice... Touch... kind of like interacting with it and it interacting back."
对柔软和可触摸覆盖物的关注为参与者最需要的感知能力提供了背景，即机器人能够处理触摸 。LFP6 解释说："有东西可以......说话、拥抱和触摸，那将会很好......触摸......就像与它互动，它也会互动回来。

SFP7 also describe feedback through touch, "I would want them to have the touch sensor and respond to me petting them because in my mind, this is like a service animal without the actual animal component."
SFP7 还描述了通过触摸获得反馈的情况，"我希望它们有触摸传感器，并能对我抚摸它们做出反应，因为在我看来，这就像一只没有动物成分的服务动物"。

Other than touch, participants indicated wanting the robot to be able to have some form of visual sensor , but not a camera . The robot being able to register that someone is near it or making certain movements was of particular interest, and participants had multiple ideas of how to achieve this, such as some sort of temperature reading or infrared. SFP2 explained, "For me, if there is some sort of IR sensor or something where we don’t have to have a camera, it could maybe sense if there is some sort of presence and can move accordingly."
除了触摸之外，与会者还表示希望机器人能够具有某种形式的视觉传感器 ，但不是摄像头 。与会者对机器人能够记录有人靠近它或做出某些动作特别感兴趣，并就如何实现这一目标提出了多种想法，例如某种温度读数或红外线。SFP2 解释说："对我来说，如果有某种红外传感器或其他东西，我们就不需要摄像头了，它或许可以感知是否有人存在，并做出相应的动作。

Sensors that could detect physiological changes were not presented to participants as a part of the sensor list, but were often requested by the participants . This included aspects such as breathing or heart rate monitors. LFP5 said, "I think that’s how I feel like the senses that would would pick up on physiology of the person. But kind of make a camera not so necessary in terms of picking up on mood." LFP1 also explained using these specific sensors may be used for reflection with their care team: "For example, I can see it was kind of like... I had... a blood pressure spike and then little Therabot recorded it and sent it to my doctor or my therapist and then they were like, P1, why did your blood pressure spike on this day? And it’s like I saw a spider or something."
可以检测生理变化的传感器并没有作为传感器清单的一部分提供给参与者，但参与者通常会提出要求 。这包括呼吸或心率监测器等方面。LFP5 说："我认为这就是我所感觉到的能够捕捉到人的生理特征的感官。但从捕捉情绪的角度来看，摄像头就没有那么必要了"。LFP1 还解释说，使用这些特定的传感器可用于与护理团队进行反思："例如，我可以看到它有点像......我......血压飙升，然后小 Therabot 记录下来并发送给我的医生或治疗师，然后他们就会问，P1，为什么你的血压会在这一天飙升？就好像我看到了蜘蛛什么的"

Sound sensors were presented to all participants, but overall their main interest in sound sensors was based upon the robot being able to recognize its name as given by the participant, or keywords 5), less focused on continuous recordings.
向所有参与者展示了声音传感器，但总体而言，他们对声音传感器的主要兴趣在于机器人能否识别参与者给出的名称或关键字 5），较少关注连续记录。

4.3.2 Robot Sounds and Behaviors. Participants were asked, once they had decided what they would like their robot to look like, to describe what sounds and behaviors the robot would be able to exhibit. Matching with the previously mentioned natural coloring, natural sounds (such as purring if they chose a cat) were requested . SFP6 gave several examples of natural sounds: "Okay, well, the cat should definitely purr you can’t have a cat that doesn’t purr... And some sort of like a light meow you don’t want to making some crying thing but just some little, you know, how they kind of they almost kind of chirp a little bit when they get really happy."
4.3.2 机器人的声音和行为。一旦参与者决定了他们希望自己的机器人是什么样子，他们就会被要求描述机器人能够发出什么样的声音和做出什么样的行为。与前面提到的自然色彩相匹配，他们还要求机器人发出自然的声音（例如，如果他们选择的是一只猫，就会发出呼噜声） 。SFP6 列举了几个自然声音的例子：SFP6举了几个自然声音的例子："好吧，猫肯定要发出呼噜声，你不能养一只不发出呼噜声的猫......还有一些像轻微的喵喵声，你不希望发出一些哭声，而只是一些小声音，你知道，当它们非常高兴时，它们会发出一些啾啾声。

We also asked whether participants desired the robot to have the potential of a human voice, which many participants indicated finding to be negative . SFP9 explained,"If it’s just, you know, kind of just being emotional support, maybe it’s better that it not say anything you would expect a human to say. Just you know, animal noises."
我们还询问了参与者是否希望机器人具有人类声音的潜能，许多参与者都表示这一点是负面的 。SFP9 解释说："如果它只是一种情感支持，也许它最好不要说出任何你希望人类说的话。你知道的，就是动物的声音。" SFP9 解释说。

Participants expressed interest in a heartbeat that they could feel and hear to calm themselves. SFP4 said the heartbeat may be useful to enact therapeutic techniques in this way, "It just came to me but like a guided meditation, sort of thing would be really cool. And so a heartbeat could be really cool there."
参与者 表示，他们对可以感受和听到的心跳声很感兴趣，因为这样可以让自己平静下来。SFP4 说，心跳可能有助于以这种方式实施治疗技术，"我只是突然想到，但像引导冥想之类的东西真的很酷。因此，心跳可能会非常酷。

Physical behaviors requested by participants followed the pattern of being more natural , such as lifting the head when pet or being able to move its tail to express itself. SFP6 gave a specific example related to the tail of the cat design: "When cats are happy, their tails kind of sway and swish a little bit, they’ll wrap their tail around their bodies or even over your arm or your leg or something."
参与者所要求的身体行为遵循了更加自然的模式 ，例如在抚摸时抬起头，或者能够移动尾巴来表达自己。SFP6 举了一个与猫尾巴设计有关的具体例子："当猫高兴的时候，它们的尾巴会摇摆和摆动一下，它们会用尾巴缠住自己的身体，甚至缠住你的胳膊或腿什么的。

However, the robot being able to express certain therapy exercises such as deep breathing or grounding was a particular behavior participants reported as potentially useful during episodes of distress, such as experiencing ruminating thoughts. LFP1 explained, "Maybe a reminder some people get stuck in like spirals, like depressive just like that and maybe like Mr. Fluffy needing to be fed might like help you get out of that." SFP2 gave a more general example of this, "I think in terms of that having a subtle sound where the person can follow that sounds and practice the exercises."
不过，机器人能够表达某些治疗练习，例如深呼吸或接地 是一种特殊的行为，据参与者报告，这种行为在痛苦发作时可能会有用，例如出现反刍的想法。LFP1 解释说："也许提醒会让一些人陷入像螺旋一样的困境，就像抑郁一样，也许像毛毛先生需要喂食一样，可能会帮助你摆脱这种困境。SFP2 举了一个更普遍的例子："我认为，如果有一种微妙的声音，人们就可以跟随这种声音进行练习。

4.3.3 Robot Connected App Requests. Two main points of interest emerged when discussing the robot connected phone application that could pair with Therabot . The first was that the app could have some sort of journal or mood tracking aspect, so that the user could look back at certain dates and get context on their emotions . LFP7 said,"So here if you were going to go to log your day, you could log how you were doing emotionally, like a little journal, maybe jot some things down, just the general sort of check in. If you want to see your data, you could see all of your previous logs, but you can also see them sort of juxtaposed with the information that Therabot itself is picking up."
4.3.3 机器人连接应用程序请求。在讨论可与 Therabot 配对的机器人连接手机应用时，出现了两个主要兴趣点 。首先，该应用程序可以具有某种日志或情绪跟踪功能，这样用户就可以回顾某些日期，了解他们的情绪背景 。LFP7 说："因此，在这里，如果您要记录一天的活动，您可以记录自己的情绪表现，就像写日记一样，也许可以记下一些事情，只是一般的签到。如果您想查看您的数据，您可以查看您之前的所有日志，但您也可以将它们与 Therabot 本身获取的信息并列起来查看。

When discussing the app, the other interest of participants was the app presenting them with a virtual avatar (figure 3) for their robot that they could access when not physically with the robot, or to further interact with the robot in the way of caring for it .
在讨论该应用程序时，参与者的另一个兴趣点是该应用程序为他们的机器人提供了一个虚拟化身（图 3），他们可以在不与机器人实际接触时访问该虚拟化身，或者以照顾机器人的方式与机器人进一步互动 。

Figure 3: Participant App Sketch and UI concepts based on participant ideas related to the connected app design.
图 3：根据与互联应用程序设计相关的参与者想法绘制的参与者应用程序草图和用户界面概念。

SFP4 related to this to a familiar digital pet, "(Have) the avatar be like a Tamagotchi sort of thing. Where like you log that you drink water or eat food or exercise or whatever, and then your avatar gets it and like so you’re taking care of the avatar creature." LFP7 gave a specific example of when they might depend on the virtual version of the robot, "Maybe I’m at the dentist office. I’m afraid of the dentist, so I’m getting stressed out. I open the Therabot app and I maybe log that I’m feeling afraid. And I do the little mindfulness exercise thing."
SFP4 将其与我们熟悉的数字宠物联系起来，"（让）化身就像一个 Tamagotchi 类似的东西。比如你记录自己喝水、吃东西、运动或其他什么的，然后你的化身就会收到，这样你就在照顾化身生物了"。LFP7 举了一个具体的例子，说明他们什么时候可能会依赖虚拟版机器人："也许我在牙医诊所。我害怕看牙医，所以压力很大。我打开 Therabot 应用程序，也许会记录下我感到害怕。然后我就会做一些正念练习"。

LFP1 gave a specific example that inspired the UI mockups of the Therabot app,"I think that might add a little bit of fun to it, like maybe each cat or each dog like certain foods and maybe each day they want a different food and that’ll just give you extra like stimulus and be like all right And you go on your phone and you click on the food they want and then they’re happy."
LFP1 举例说明了 Therabot 应用的用户界面模型的灵感来源，"我认为这可能会为它增添一些乐趣、也许每只猫或狗都喜欢某些食物，也许它们每天都想吃不同的食物，这样就会给你带来额外的刺激，然后你就可以在手机上点击它们想吃的食物，然后它们就会很开心。"

The ability to do therapy exercises through the app was also of interest to participants.
通过应用程序 进行治疗练习的功能也引起了参与者的兴趣。

4.3.4 Robot and App Privacy Concerns. Participants were also asked about any privacy concerns that they may have regarding data collected by the robot’s sensors. Overall participants felt comfortable sharing the data that was collected with their care team, specifically their therapists . LFP7 indicated believing the robot would be connected to their care team regardless, "If this is like integrated to therapy, I would imagine that therapists would be given that information. But probably there would be like a consent thing to that:"
4.3.4机器人和应用程序的隐私问题。与会者还被问及他们对机器人传感器收集的数据是否有隐私方面的顾虑。总的来说，参与者认为与护理团队（尤其是治疗师）共享收集到的数据是件很舒服的事 。LFP7 表示，无论如何，机器人都会与他们的护理团队建立联系，"如果这与治疗相结合，我想治疗师会得到这些信息。但这可能需要征得同意：" LFP7 说。

However, being able to control what information was sent and when was critical for participants to retain the feeling of control in regard to data collection . Such as SFP3,"I think it would be good to know where the information is going to. I think personally I’d only really want it between me and my therapists and care team, but I would understand like some people are a little bit closer with their family and would also like share it with them."
然而，能够控制信息的发送内容和时间对于参与者保持对数据收集的控制感至关重要 。例如 SFP3，"我认为知道信息将被发送到哪里是件好事。我个人认为，我只想在我和我的治疗师及护理团队之间分享这些信息，但我理解有些人与家人的关系更密切一些，他们也希望与家人分享这些信息。

LFP1 provided a specific reason to why they felt it was necessary to monitor who had access to their information as it was collected by the robot. LFP1 said, "I’ve had situations in the past where my mental illness was used as a way to like sort of make me feel show me as unreliable during a situation and I was just thinking of a situation where there was some type of something like a custody battle over a kid where maybe one spouse would take the robot and be like look my wife is crazy look at the dog. It like records all."
LFP1 提供了一个具体的理由，说明为什么他们认为有必要在机器人收集他们的信息时，监控谁可以访问这些信息。LFP1 说："我过去曾遇到过这样的情况，我的精神疾病被用来让我觉得自己在某种情况下是不可靠的，我只是在想，在某种情况下，比如孩子的监护权争夺战，也许夫妻中的一方会拿着机器人，说我妻子疯了，看看这只狗。它就像所有的记录"。

4.4 Workshop Differences
4.4 工作坊的差异

There were some differences within participant responses between the two workshops, perhaps related to the amount of time allotted to consider each of the different aspects of the robot and its connected app. In the following sections " " indicates that the number of participant responses are from only the long form workshop, while " " remains the total number of responses from all 15 participants.
style="font-family: "Times New Roman" min-height: 11pt; font-size: 11pt;" data-imersive-translate-walked="287aa928-15d3-4c1f-b165-c43cec8e695d">在两个研讨会中，与会者的回答存在一些差异，这可能与分配给考虑机器人及其连接应用程序不同方面的时间有关。在以下章节中，" "表示参与者的回复数量仅来自长篇研讨会，而" n "则表示所有 15 位参与者的回复总数。

One aspect that participants from the long form workshops were more interested in was the robot connected app being able to guide them through therapeutic techniques that they had developed with their care team . LFP1 gave a specific example,"But it’s like remember the grounding technique, what are five things you can see, four things you can smell and it’s just like that kind of thing, you know, but it’s with your Therabot so."
长篇研讨会的参与者更感兴趣的一个方面是，与机器人连接的应用程序能够指导他们完成与护理团队共同开发的治疗技巧 。LFP1 举了一个具体的例子："但这就像记住接地技术一样，你能看到的五样东西是什么，你能闻到的四样东西是什么，它就像那种东西，你知道，但它是与你的 Therabot 一起使用的。

Notably, it was the participants from the short workshops, who were not affiliated with Centerstone, that were most interested in being able to send information to their Therapist. While the long form participants mentioned being comfortable sharing the data collected by the robot and app, nine participants from the short form workshop were comfortable passing along that information.
值得注意的是，与中心石无关的短期研讨会参与者对能够向治疗师发送信息最感兴趣。长式研讨班的参与者提到他们愿意分享机器人和应用程序收集到的数据，而短式研讨班的九名参与者则愿意将这些信息传递给治疗师。

Two major differences between the workshop forms was that in the long form workshop participants were asked where they would use the robot the most, which was not included in the short form workshop due to time. Overall participants indicated wanting the robot in the bedroom and the living room .
研讨会形式之间的两个主要区别是，在长形式研讨会上，与会者被问及他们最常在哪里使用机器人，而在短形式研讨会上，由于时间关系，没有问到这个问题。总的来说，与会者表示希望机器人放在卧室 和客厅 。

During the short form workshops, participants were asked if they had any concerns over using the robot overall, rather than having the focus be to voice concerns along the way such as in the long form workshop. Participants in the short form workshop voiced concerns over users becoming attached to the robot . Despite these differences, the two workshops engaged the participants in similar ways, and in most areas participants relayed similar ideals across the methods.
在简式研讨会上，与会者被问及他们对使用机器人的整体情况是否有任何顾虑，而不是像在长式研讨会上那样，重点是提出使用过程中的顾虑。简短研讨会的与会者对用户对机器人产生依赖表示担忧 。尽管存在这些差异，但两场研讨会还是以类似的方式吸引了与会者，而且在大多数领域，与会者都表达了类似的理想。

4.5 Clinician Workshops
4.5 临床医生讲习班

Although only four clinicians participated in the workshops, there were themes that emerged based upon what aspects of the robot and the connected phone app they felt would be beneficial for their clients, those living with depression. In the following section " " indicates a clinician participant.
虽然只有四位临床医生参加了研讨会，但他们认为机器人和联网手机应用的哪些方面对他们的客户（即抑郁症患者）有益，因此出现了一些主题。在下面的章节中，" "表示一位临床医生参与者。

4.5.1 Clinician Robot and App Requests. While clinicians overall did not seem to have a physical design they felt would be best for their clients, they did express strong feelings on certain sensors. Physiological sensors were the most common, with all four clinicians indicating wanting some sort of sensor that could read clients’ body signals. Heart rate monitoring was the most specifically requested , with breathing patterns being important as well . With Clinician 4 speaking on heart rate specifically,"I’m thinking about the ways that most of the folks that I work with have like co-morbid concerns with depression and anxiety. So I thought okay increase in heart rate might be an indicator that a panic attack is coming."
4.5.1临床医生对机器人和应用程序的要求。虽然临床医生总体上似乎没有他们认为最适合客户的物理设计，但他们确实对某些传感器表达了强烈的情感。生理传感器是最常见的，所有四位临床医生都表示希望使用某种可以读取客户身体信号的传感器。心率监测是最具体的要求 ，呼吸模式也很重要 。临床医生 4 在谈到心率时特别指出："我在想，与我一起工作的大多数人都有抑郁和焦虑等并发症。因此，我认为心率增加可能是恐慌症即将发作的一个指标。

Clinicians provided specific examples of when a sound sensor may be useful for the robot to respond when the participant was feeling specific emotions . C3 said,"Audio sensor because you could pick up on like people sighing."
临床医生提供了具体的例子，说明声音传感器在什么情况下可以帮助机器人在受试者感受到特定情绪时做出反应 。C3 说："声音传感器可以捕捉到人们的叹息声。

Cameras were not as important to clinicians as audio sensors, as they felt that a camera being present may actually provide a barrier to the adoption of the robot by their clients . C1 gave an example of why this barrier may exist,"I just think again of my folks, if they have any kind of paranoia, you know, I have people that just will tape up the camera even on their cell phone. I think that could be a barrier for some folks."
对于临床医生来说，摄像头并不像音频传感器那样重要，因为他们认为摄像头的存在实际上可能会阻碍客户采用机器人 。C1 举例说明了这一障碍存在的原因："我只是再次想到我的客户，如果他们有任何偏执狂，你知道，我的客户甚至会用胶带将手机摄像头粘起来。我认为这可能会成为某些人的障碍。

One potential capability of the robot all clinicians expressed interested in was the ability to place calls from the robot. Specifically for calling for aid, either from emergency services or the clients support system . C2 gave an example of a client needing aid, "Yes, to be able to call emergency services is if necessary, you know, because like if we have a client who is struggling with diabetes and taking their meds, sugar bottoms out, they can feel it, but not going to have enough time, they can yell out, hey, send help."
所有临床医生都对机器人的一项潜在功能感兴趣，那就是通过机器人拨打电话。具体来说，就是呼叫紧急服务或客户支持系统提供援助 。C2 举了一个客户需要帮助的例子："是的，如果有必要，可以呼叫紧急服务，你知道，因为如果我们有一个客户正在与糖尿病作斗争，正在吃药，血糖降到了谷底，他们能感觉到，但没有足够的时间，他们可以大喊一声：嘿，快来帮忙！" 。

For the robot connected app, clinicians believed the app could be a way of journaling or tracking information regarding certain experiences or feelings for their clients . C4 exampled,"The app should allow them to provide context if they need to like to write down like, maybe their numbers show that they had a panic attack and like it’s not going to really show like what was actually going on."
对于机器人连接应用程序，临床医生认为，该应用程序可以成为客户记录或跟踪有关某些经历或感受的信息的一种方式 。C4举例说："如果他们需要写下类似的内容，应用程序应该允许他们提供相关信息，例如，也许他们的数字显示他们恐慌发作了，但这并不能真正显示实际发生了什么。

An aspect that carried both between the robot and the app was the ability for clinicians clients to be able to set things such as reminders , specifically reminders about tools that the clients learned during therapy sessions . C1 said,"Like, like maybe they do like a daily breathing exercise where they have to like... there’s some form of in and out like almost breathing that they sync their deep breathing with the Therabot and that’s how they do like a mindfulness - like maybe deep breathing check in and they you know when the when the robot expands they breathe."
机器人和应用程序之间的一个共同点是，临床医生可以为客户设置提醒事项 ，特别是提醒客户在治疗过程中学到的工具 。C1 说："比如，也许他们每天都会进行呼吸练习，他们必须像......有某种形式的进气和出气，就像几乎呼吸一样，他们将自己的深呼吸与 Therabot 同步，他们就是这样进行正念--就像也许是深呼吸检查，当机器人扩张时，他们就会呼吸。

Overall, clinicians were interested in receiving the data from their clients and considered that the information from this data may aid in identifying if specific clients needed more or different care. C4 indicated the data could potentially impact what interventions they use, "I was gonna say that it could help us better understand if what we’re doing is working with them. So, like, we increased their sessions and maybe it didn’t help. Like, maybe that was the wrong intervention. You know, maybe there were things that we were doing that was actually like, you know, triggering them in a way that wasn’t beneficial to them or something like that." Or, perhaps the data collected could show that changes were occurring through the ongoing therapy, such as C3, "I feel like a lot of my clients are just consistently like, yeah, I don’t feel like anything is getting better. But it’d be nice to have like some sort of like actual evidence that like, well, you know, this particular symptoms decreased over the past like month or something."
总的来说，临床医生对接收来自客户的数据很感兴趣 ，并认为这些数据信息可以帮助确定特定客户是否需要更多或不同的护理。C4 表示，这些数据可能会影响他们所使用的干预措施，"我想说的是，这可以帮助我们更好地了解我们所做的是否对他们有效。比如说，我们增加了他们的疗程，但也许并没有什么用。也许那是错误的干预。你知道，也许我们所做的某些事情，实际上就像，你知道，以一种对他们无益的方式触发了他们，或者类似的事情"。或者，也许收集到的数据可以表明，正在进行的治疗正在发生变化，比如 C3，"我觉得我的很多客户只是一直在说，是啊，我不觉得有什么好转。但如果能有一些实际证据证明，比如，你知道，在过去的一个月里，这种特殊症状有所减轻，那就更好了"。

Privacy concerns extended beyond the data collected by the robot about clients and to clinicians themselves when interacting with their clients that have the robot. Three clinicians mentioned specific privacy concerns, and two specifically mention their own privacy. C3 touched on this in reference to using the robot in therapy sessions "I don’t want to be recorded... And I’d be worried if there would be any way where they could like take something you said out of context and be like look at this like 2 second portion of what C3 said, you know, so it would just make me uncomfortable and I think I would be not on my best ability."
对隐私的担忧不仅限于机器人收集的客户数据，还包括临床医生本身在与使用机器人的客户互动时对隐私的担忧。三位临床医生提到了具体的隐私问题，其中两位特别提到了他们自己的隐私。C3 在谈到在治疗过程中使用机器人时提到："我不想被记录......我担心他们会断章取义，把你说过的话断章取义，比如看看 C3 说过的 2 秒钟的部分内容，你知道，这样会让我很不舒服，我觉得我的能力没有发挥到极致"。

The robot having other capabilities were more divisive, such as the robot having the ability to speak, with one clinician out right saying that it would make their clients nervous, while another thought it could be beneficial for indicating reminders and similar information to clients.
对于机器人是否具备其他功能，医生们的意见分歧较大，比如机器人是否能说话，一位医生直言这会让客户感到紧张，而另一位医生则认为这有利于向客户发送提醒和类似信息。

4.5.2 Clinician Similarities to Clients. While clinicians and those living with depression were approaching the robot and app design from different angles, there was overlap in some of the requests. One of the most clear requests by both populations was the robot being able to detect bodily changes in the user, through some sort of physiological sensor .
4.5.2临床医生与客户的相似之处。尽管临床医生和抑郁症患者从不同的角度来考虑机器人和应用程序的设计，但在某些要求上还是存在重叠。这两类人群最明确的要求之一就是机器人能够通过某种生理传感器来检测用户的身体变化 。

Both clinicians and those living with depression indicated that including cameras on the robot may hinder its ability to be used in a therapeutic setting. Clinicians felt that it may adversely effect their clients , while those living with depression felt that it could wouldn’t be necessary , particularly if there were alternatives presented.
临床医生和抑郁症患者都表示，在机器人上安装摄像头可能会妨碍其在治疗环境中的应用。临床医生认为，这可能会对他们的客户产生不利影响 ，而抑郁症患者则认为没有必要 ，尤其是在有替代方案的情况下。

For the robot connected app, tracking through journaling, such as a mood journal or event journal, was felt to be helpful for reflection and notification both by clinicians and those living with depression . The data collected both by the robot and the app was seen as potentially useful for clinicians , and overall those living with depression felt comfortable sharing that information with their therapists .
对于与机器人连接的应用程序，临床医生 和抑郁症患者 认为通过日志（如心情日志或事件日志）进行跟踪有助于反思和通知。机器人和应用程序收集的数据被认为对临床医生 有潜在的帮助，而且总体而言，抑郁症患者认为与他们的治疗师分享这些信息很舒服 . 。

5 DISCUSSION
5 讨论

While most often discussed in relation to children, the importance of touch on well-being is well documented for reducing cortisol levels and blood pressure [17]. Touch, and the softness of the robot to encourage touch, was important for those living with depression, perhaps for a similar reason. Participants preferred robots that more closely resembled pets (cats or dogs) and had more natural colors, which is consistent with traditional animal-assisted therapy. While dogs are the most common animal assisted therapy animal [32], participants overall indicated having or have had in the past, cats. Cats, while not as common, have also been used with those experiencing illnesses such as depression [45]. The desire for robots of different animal types indicates the importance of a modular and adaptable platform that can take on the characteristics of different animals by altering its appearance and behaviors.
虽然触觉对儿童的影响最常被讨论，但触觉对于降低皮质醇水平和血压的重要性也是有据可查的[17]。也许出于类似的原因，触摸以及机器人的柔软度对抑郁症患者来说非常重要。参与者更喜欢与宠物（猫或狗）更相似、颜色更自然的机器人，这与传统的动物辅助疗法是一致的。虽然狗是最常见的动物辅助治疗动物[32]，但总的来说，参与者表示他们过去养过或曾经养过猫。猫虽然并不常见，但也被用于治疗抑郁症等疾病[45]。对不同动物类型机器人的渴望表明，一个模块化、适应性强的平台非常重要，它可以通过改变外观和行为来呈现不同动物的特征。

Participants in all workshops identified features that might best be provided by a companion mobile application. Aside from providing a reliable configuration interface and settings hub for the robotic companion, an app has the potential to enhance the therapy process. An app is well-suited to serve as a link to the patient’s comforting companion and simultaneously provide features that have been shown to improve therapeutic outcomes. For example, journaling, which has recognized therapeutic benefits [47], was underscored by both patients and clinicians as a valuable feature to integrate into a companion application. Combined with positive reinforcement from the patient’s relationship with their robot, app-based incentives may lead to an increase in consistency of completing check-in, journal entries, and introspection activities. Regarding daily logging, clinicians indicated value in patients charting their moods and medication for review over time, as this would provide useful insights when assessing potential trends and treatment outcomes. In situations where patients experience heightened anxiety or emotional distress, an app can provide an avenue for them to document their experiences, detailing, for example, context, their primary symptoms, and duration, which would be a helpful reference in subsequent therapy sessions.
所有研讨会的与会者都指出了最好由配套移动应用程序提供的功能。除了为机器人伴侣提供可靠的配置界面和设置中心外，应用程序还有可能改善治疗过程。应用程序非常适合作为连接病人的舒适伴侣的纽带，同时还能提供已证明能改善治疗效果的功能。例如，日记具有公认的治疗效果[47]，患者和临床医生都强调日记是一种有价值的功能，可以整合到伴侣应用程序中。结合患者与机器人关系的积极强化，基于应用程序的激励措施可能会提高患者完成签到、日记和反省活动的一致性。关于每日记录，临床医生表示，患者将自己的情绪和用药情况制成图表供长期审查很有价值，因为这将为评估潜在趋势和治疗效果提供有用的见解。在患者经历高度焦虑或情绪困扰的情况下，应用程序可以为他们提供记录其经历的途径，例如详细记录背景、主要症状和持续时间，这将为后续治疗提供有用的参考。

In the app, options for tailoring the robot’s virtual aesthetic may foster a sense of ownership and connection. Additionally, the introduction of gamified elements could enhance user engagement, creating a mutual care dynamic between the patient and the robot. For example, when a patient logs their food and water intake, they could simultaneously address the virtual companion’s needs. Linked reminder systems between the physical robot and its virtual counterpart could be established: if a patient neglects to log their meals, the robot could alert them by exhibiting signs of hunger. This not only serves as a prompt for self-care for the user but also reinforces a sense of responsibility towards their robotic companion, leveraging the concept of the "helper’s high" [10]. Given that clinicians might employ the robot for diverse age groups, the application should cater to varying user preferences. For younger demographics, integrating augmented reality (AR) capabilities might be effective.
在应用程序中，定制机器人虚拟美学的选项可以培养用户的主人翁意识和联系感。此外，游戏化元素的引入可以提高用户的参与度，在病人和机器人之间形成一种相互关爱的动态关系。例如，当病人记录食物和水的摄入量时，他们可以同时满足虚拟伴侣的需求。可以在实体机器人和虚拟机器人之间建立关联提醒系统：如果病人忽略了记录自己的进餐情况，机器人可以通过显示饥饿迹象来提醒他们。这不仅是对用户自我保健的一种提示，还能利用 "帮助者的高涨"[10]这一概念，强化他们对机器人伴侣的责任感。鉴于临床医生可能会将机器人用于不同年龄段的人群，因此应用程序应迎合不同用户的偏好。对于年轻人群，集成增强现实（AR）功能可能会很有效。

Clinicians and patients also indicated that implementing interactive exercises utilizing techniques learned in therapy sessions, such as meditation and mindfulness, would be beneficial. Considering its form factor and weight, Therabot , when situated in a patient’s lap, could serve as an excellent grounding element, providing tactile, auditory, and visual stimuli; furthermore, the robot could perform soothing haptic vibrations, adaptive throughout the duration of the guided session, configured to mimic purring, breathing. These potential uses highlight the importance of a physically embodied robot and software application working together rather than choosing one or the other.
临床医生和患者还表示，利用在治疗过程中学到的技巧（如冥想和正念）进行互动练习将大有裨益。考虑到其外形尺寸和重量，Therabot 放在患者腿上时，可以作为一个极佳的接地元素，提供触觉、听觉和视觉刺激；此外，机器人还可以在整个指导过程中进行舒缓的触觉振动，并模仿呼噜声和呼吸声。这些潜在用途凸显了实体机器人和软件应用程序协同工作的重要性，而不是二选一。

The insights and preferences presented by participants must also be evaluated in the context of technical feasibility. Although the robot’s current implementation has the exterior appearance of a dog, additional exterior coverings including those representing a cat have been prototyped and are planned to be used in upcoming design and evaluation efforts. Furthermore, the robot’s underlying structure is designed in a modular fashion that supports exchanging the dog tail-wagging module for a more articulated cat tail module. We are currently conducting user evaluations of a generative audio subsystem and expanded haptic capabilities (see [5] for more details) that provide support for additional types of animal the robot can exemplify.
此外，还必须根据技术可行性对参与者提出的见解和偏好进行评估。虽然该机器人目前的外形是一只狗，但包括猫在内的其他外部覆盖物已经完成原型设计，并计划在接下来的设计和评估工作中使用。此外，机器人的底层结构采用模块化设计，支持将狗摇尾模块更换为铰接性更强的猫尾模块。目前，我们正在对生成音频子系统和扩展触觉功能（详见文献[5]）进行用户评估，这些功能可为机器人示范更多类型的动物提供支持。

As most pets and support animals are able to demonstrate a multi-modal understanding of their environment, it is intuitive that participants expressed a desire for the robot to be able to recognize their presence and location in the environment. However, participants also indicated that use of a camera system creates substantial privacy concerns. In order to support a more privacy-conscious method for understanding the locations of people in the robot’s environment, we are beginning technical work evaluating a variety of sensing solutions that do not include cameras. One promising option is the use of ultra-wideband (UWB) radar to understand the environment. For example, UWB-based solutions have been deployed successfully as a privacy sensitive way of monitoring patients in assisted living facilities [34]. Achieving a small form factor that can be contained on the robotic platform remains a challenge and is an area of ongoing technical development.
由于大多数宠物和辅助动物都能表现出对环境的多模态理解，因此参与者直观地表示希望机器人能够识别它们在环境中的存在和位置。不过，与会者也表示，使用摄像系统会产生很大的隐私问题。为了支持一种更注重隐私的方法来了解机器人环境中人的位置，我们正在开始技术工作，评估各种不包括摄像头的传感解决方案。使用超宽带 (UWB) 雷达了解环境是一个很有前景的方案。例如，基于 UWB 的解决方案已成功应用于生活辅助设施中对隐私敏感的病人监控[34]。实现可容纳在机器人平台上的小型外形仍然是一个挑战，也是一个正在进行技术开发的领域。

Participants also expressed a desire for the robot to be able to measure their physiological state (e.g., respiration or heart rate) and take appropriate actions. We are currently conducting studies evaluating the integration of a wearable PPG heart sensor, similar to those included in most smart watches and fitness trackers, to adapt the robot’s simulated heartbeat in an effort to induce a helpful physiological change in the user. While common wearable consumer physiological sensors are a useful avenue towards understanding the user’s state, UWB radar is also capable of tracking the respiration and heart rate of people without the need for direct contact. For example, commercially available UWB bedside units have demonstrated the ability to track heart rate, respiration, and sleep quality with clinically useful levels of accuracy .
参与者还表示希望机器人能够测量他们的生理状态（如呼吸或心率），并采取适当的行动。目前，我们正在开展研究，评估如何整合可穿戴 PPG 心脏传感器（类似于大多数智能手表和健身追踪器中的传感器），以调整机器人的模拟心跳，从而诱导用户产生有益的生理变化。虽然常见的可穿戴消费类生理传感器是了解用户状态的有用途径，但 UWB 雷达也能在无需直接接触的情况下跟踪人的呼吸和心率。例如，市场上销售的 UWB 床边设备已证明能够跟踪心率、呼吸和睡眠质量，其精确度达到了临床有用的水平 . 。

As our ultimate focus is on deploying a personalized robot in the homes of adults living with depression, future work will involve studying the dynamics of interactions with the robot and companion application over longer periods of time.
由于我们的最终目标是在患有抑郁症的成年人家中部署个性化机器人，因此未来的工作将包括研究在较长时间内与机器人和伴侣应用的互动动态。

6 CONCLUSION
6 结论

Overall, after the completion of the workshops, participants provided guidelines to design a socially assistive robot that they felt would aid in managing the symptoms of depression. Future work will continue to delve deeper into specific aspects of continued development of Therabot for this population, including further exploration into specific sensors and behaviors for supporting the user. This research supports the design of a personalizable Therabot and future user evaluation studies of the adaptable robot in the homes of those living with depression, to test the efficacy of this type of device on the management of depression symptoms.
总之，在研讨会结束后，与会者为设计一款社交辅助机器人提供了指导，他们认为这款机器人有助于控制抑郁症状。未来的工作将继续深入探讨为这一人群继续开发 Therabot 的具体方面，包括进一步探索支持用户的特定传感器和行为。这项研究支持设计个性化的Therabot ，并支持未来在抑郁症患者家中对可适应的机器人进行用户评估研究，以测试这类设备对控制抑郁症状的功效。

ACKNOWLEDGMENTS
致谢

A special thanks to Alexis King for her time working with the transcriptions and Audrey Aldridge for her sketches.
在此特别感谢亚历克西斯-金（Alexis King）在抄写过程中付出的时间，以及奥德丽-奥尔德里奇（Audrey Aldridge）的素描。

REFERENCES
参考文献

[1] Muneeb Imtiaz Ahmad, Omar Mubin, and Joanne Orlando. 2017. A systematic review of adaptivity in human-robot interaction. Multimodal Technologies and Interaction 1, 3 (2017), 14. Caldeira de Melo, Samila Sathler Tavares Batistoni, Meire Cachioni, and Patrick CK Hung. 2022. Effects of social robots on depressive symptoms in older adults: a scoping review. Library hi tech 40, 5 (2022), 1108-1126.
[1] Muneeb Imtiaz Ahmad、Omar Mubin 和 Joanne Orlando。2017.人机交互中的适应性系统综述》。多模态技术与交互》1，3（2017），14。Caldeira de Melo、Samila Sathler Tavares Batistoni、Meire Cachioni 和 Patrick CK Hung。2022.社交机器人对老年人抑郁症状的影响：范围综述》。Library hi tech 40, 5 (2022), 1108-1126.

[3] Casey C. Bennett, Selma Sabanovic, Jennifer A. Piatt, Shinichi Nagata, Lori Eldridge, and Natasha Randall. 2017. A Robot a Day Keeps the Blues Away. In 2017 IEEE International Conference on Healthcare Informatics (ICHI). IEEE, Utah, USA, 536-540. https://doi.org/10.1109/ICHI.2017.43
[3] Casey C. Bennett、Selma Sabanovic、Jennifer A. Piatt、Shinichi Nagata、Lori Eldridge 和 Natasha Randall.2017.一天一个机器人，远离忧郁症。In 2017 IEEE International Conference on Healthcare Informatics (ICHI).IEEE, Utah, USA, 536-540. https://doi.org/10.1109/ICHI.2017.43

[4] Cindy L. Bethel, Zachary Henkel, Sarah Darrow, and Kenna Baugus. 2018. on Digital Intelligence for Systems and Machines (DISA). IEEE, Kosice, Slovakia, 23-30. https://doi.org/10.1109/disa.2018.8490642
[4] Cindy L. Bethel、Zachary Henkel、Sarah Darrow 和 Kenna Baugus.2018. 系统和机器的数字智能（DISA）。IEEE, Kosice, Slovakia, 23-30. https://doi.org/10.1109/disa.2018.8490642

[5] Cindy L. Bethel, Zachary Henkel, Kenna Henkel, Jade Thompson, and Kyler Smith. 2023. From Components to Caring: The Development Trajectory of a Socially Therapeutic Assistive Robot (STAR) Named Therabot . In 2023 World Symposium on Digital Intelligence for Systems and Machines (DISA). 31-45. https: //doi.org/10.1109/DISA59116.2023.10308910
[5] Cindy L. Bethel、Zachary Henkel、Kenna Henkel、Jade Thompson 和 Kyler Smith。2023.从组件到关爱：名为 Therabot 的社会治疗辅助机器人 (STAR) 的发展轨迹 .In 2023 World Symposium on Digital Intelligence for Systems and Machines (DISA).31-45. https: //doi.org/10.1109/DISA59116.2023.10308910

[6] Buddy Robot Video 2015. BUDDY: Your Family’s Companion Robot. Retrieved September 28, 2023 from https://www.youtube.com/watch?v=51yGC3iytbY&ab_ channel Buddy
"font-family: "Times New Roman"; min-height: 11pt; font-size: 11pt;" data-imersive-translate-walked="287aa928-15d3-4c1f-b165-c43cec8e695d">[6] 巴迪机器人视频 2015。BUDDY: Your Family's Companion Robot.于 2023 年 9 月 28 日从 https://www.youtube.com/watch?v=51yGC3iytbY&ab_ channel = Buddy 检索。

[7] Cynthia K Chandler, Torey L Portrie-Bethke, Casey A Barrio Minton, Delini M Fernando, and Dana M O’Callaghan. 2010. Matching animal-assisted therapy techniques and intentions with counseling guiding theories. Journal of Mental Health Counseling 32, 4 (2010), 354-374.
[7] Cynthia K Chandler, Torey L Portrie-Bethke, Casey A Barrio Minton, Delini M Fernando, and Dana M O'Callaghan.2010.将动物辅助治疗技术和意图与心理咨询指导理论相匹配》（Matching animal-assisted therapy techniques and intentions with counseling guiding theories.心理健康咨询期刊》（Journal of Mental Health Counseling）32，4（2010），354-374。

[8] Jean-Christophe Chauvet-Gélinier, Benoit Trojak, Bénédicte Vergès-Patois, Yves Cottin, and Bernard Bonin. 2013. Review on depression and coronary heart disease. Archives of cardiovascular diseases 106, 2 (2013), 103-110.
[8] Jean-Christophe Chauvet-Gélinier、Benoit Trojak、Bénédicte Vergès-Patois、Yves Cottin 和 Bernard Bonin。2013.抑郁症与冠心病综述》。心血管疾病档案》106，2 (2013)，103-110。

[9] Shu-Chuan Chen, Wendy Moyle, Cindy Jones, and Helen Petsky. 2020. A social robot intervention on depression, loneliness, and quality of life for Taiwanese older adults in long-term care. International Psychogeriatrics 32, 8 (2020), 981-991. https://doi.org/10.1017/S1041610220000459
[9] Shu-Chuan Chen, Wendy Moyle, Cindy Jones, and Helen Petsky.2020.社会机器人对台湾长期护理老年人抑郁、孤独和生活质量的干预。国际老年精神病学》，32，8 (2020)，981-991。https://doi.org/10.1017/S1041610220000459。

[10] Chatchai Chirapornchai, Paul Bremner, and Joseph E. Daly. 2021. Helper’s High with a Robot Pet. In Companion of the 2021 ACM/IEEE International Conference on Human-Robot Interaction (Boulder, CO, USA) (HRI ’21 Companion). Association for Computing Machinery, New York, NY, USA, 229-233. https://doi.org/10.
[10] Chatchai Chirapornchai, Paul Bremner, and Joseph E. Daly.2021.2021 年 ACM/IEEE 人机交互国际会议（美国科罗拉多州博尔德市）论文集（HRI '21 Companion）。美国计算机协会，纽约州纽约市，229-233。https://doi.org/10。 。

[11] Sawyer Collins, Daniel Hicks, Zachary Henkel, Kenna Baugus Henkel, Jennifer A. Piatt, Cindy L. Bethel, and Selma Sabanovic. 2023. What Skin Is Your Robot In?. In Companion of the 2023 ACM/IEEE International Conference on Human-Robot Interaction (March 2023), 511-515.
[11] Sawyer Collins, Daniel Hicks, Zachary Henkel, Kenna Baugus Henkel, Jennifer A. Piatt, Cindy L. Bethel, and Selma Sabanovic.2023.你的机器人是什么皮肤？In Companion of the 2023 ACM/IEEE International Conference on Human-Robot Interaction (March 2023), 511-515.

[12] Sawyer Collins, Selma Šabanović, Marlena Fraune, Natasha Randall, Lori Eldridge, Jennifer A. Piatt, Casey C. Bennett, and Shinichi Nagata. 2018. Sensing Companions: Potential Clinical Uses of Robot Sensor Data for Home Care of Older Adults with Depression. In Companion of the 2018 ACM/IEEE International Conference on Human-Robot Interaction (Chicago, IL, USA) (HRI ’18). Association for Computing Machinery, New York, NY, USA, 89-90. https://doi.org/10.1145/3173386.3177047
[12] Sawyer Collins、Selma Šabanović 、Marlena Fraune、Natasha Randall、Lori Eldridge、Jennifer A. Piatt、Casey C. Bennett 和 Shinichi Nagata。2018.感知同伴：机器人传感器数据在抑郁症老年人家庭护理中的潜在临床用途》。In Companion of the 2018 ACM/IEEE International Conference on Human-Robot Interaction (Chicago, IL, USA) (HRI '18).美国计算机协会，纽约州纽约市，89-90。https://doi.org/10.1145/3173386.3177047。

[13] Sara Cooper, Alessandro Di Fava, Carlos Vivas, Luca Marchionni, and Francesco Ferro. 2020. ARI: the Social Assistive Robot and Companion. In 2020 29th IEEE International Conference on Robot and Human Interactive Communication (ROMAN). IEEE, Naples, Italy, 745-751. https://doi.org/10.1109/RO-MAN47096.2020. 9223470
[13] Sara Cooper、Alessandro Di Fava、Carlos Vivas、Luca Marchionni 和 Francesco Ferro。2020.ARI：社交辅助机器人与伴侣。2020年第29届IEEE机器人与人类交互通信国际会议（ROMAN）。IEEE，意大利那不勒斯，745-751。https://doi.org/10.1109/RO-MAN47096.2020。9223470

[14] Feil-Seifer D. and Matarić M.J. 2011. Socially assistive robotics. IEEE Robotics & Automation Magazine 18, 1 (2011), 24-31.
[14] Feil-Seifer D. and Matarić M.J. 2011.社会辅助机器人。IEEE Robotics & Automation Magazine 18, 1 (2011)，24-31.

[15] Francesca Dino, Rohola Zandie, Hojjat Abdollahi, Sarah Schoeder, and Mohammad H. Mahoor. 2019. Delivering Cognitive Behavioral Therapy Using A Conversational Social Robot. In 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 2089-2095. https://doi.org/10.1109/IROS40897.2019. 8968576
[15] Francesca Dino、Rohola Zandie、Hojjat Abdollahi、Sarah Schoeder 和 Mohammad H. Mahoor。2019.使用会话社交机器人提供认知行为疗法》。2089-2095. https://doi.org/10.1109/IROS40897.2019.8968576

[16] Diana Duong and Lauren Vogel. 2023. Overworked health workers are "past the point of exhaustion". CMAJ 195, 8 (Feb. 2023), E309-E310.
[16] Diana Duong 和 Lauren Vogel。2023.过度劳累的医护人员已 "精疲力竭"。CMAJ 195, 8 (Feb. 2023), E309-E310.

[17] Tiffany Field. 2010. Touch for socioemotional and physical well-being: A review. Developmental review 30,4 (2010), 367-383.
[17] Tiffany Field.2010.触摸促进社会情感和身体健康：A review.Developmental Review 30,4 (2010)，367-383。

[18] Eiko I. Fried and Randolph M. Nesse. 2015. Depression sum-scores don’t add up: why analyzing specific depression symptoms is essential. BMC medicine 13, 1 (2015), 1-11.
[18] Eiko I. Fried and Randolph M. Nesse.2015.抑郁症总分不相加：为什么分析特定抑郁症状至关重要？BMC medicine 13, 1 (2015), 1-11.

[19] Arzu Guneysu and Bert Amrich. 2017. Socially assistive child-robot interaction in physical exercise coaching. In 2017 26th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, Lisbon, Portugal, 670-675. https://doi.org/10.1109/roman.2017.8172375
[19] Arzu Guneysu 和 Bert Amrich。2017.体育锻炼教练中的社交辅助型儿童-机器人交互。In 2017 26th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN).IEEE, Lisbon, Portugal, 670-675. https://doi.org/10.1109/roman.2017.8172375

[20] Viviana E. Horigian, Renae D. Schmidt, and Daniel J. Feaster. 2021. Loneliness, Mental Health, and Substance Use among US Young Adults during COVID-19. Journal of Psychoactive Drugs 53, 1 (2021), 1-9. https://doi.org/10.1080/02791072. 2020.1836435 PMID: 33111650.
[20] Viviana E. Horigian, Renae D. Schmidt, and Daniel J. Feaster.2021.COVID-19 期间美国年轻人的孤独感、心理健康和药物使用》。Journal of Psychoactive Drugs 53, 1 (2021), 1-9. https://doi.org/10.1080/02791072.2020.1836435 pmid: 33111650.

[21] Melanie G. Jones, Simon M. Rice, and Susan M. Cotton. 2019. Incorporating animal-assisted therapy in mental health treatments for adolescents: A systematic review of canine assisted psychotherapy. PloS one 14, 1 (2019).
[21] Melanie G. Jones, Simon M. Rice, and Susan M. Cotton.2019.将动物辅助疗法纳入青少年心理健康治疗：犬类辅助心理疗法系统综述》。PloS one 14, 1 (2019).

[22] Nina Jøranson, Ingeborg Pedersen, Anne Marie Mork Rokstad, and Camilla Ihlebaek. 2015. Effects on symptoms of agitation and depression in persons with dementia participating in robot-assisted activity: a cluster-randomized controlled trial. Journal of the American Medical Directors Association 16, 10 (2015), 867-873.
[22] Nina Jøranson, Ingeborg Pedersen, Anne Marie Mork Rokstad, and Camilla Ihlebaek.2015.对参与机器人辅助活动的痴呆症患者躁动和抑郁症状的影响：分组随机对照试验》（Effects on symptoms of agitation and depression in persons with dementia participating in robot-assisted activity: a cluster-randomized controlled trial.美国医务主任协会期刊》16，10（2015），867-873。

[23] Sun Kang, Dong-Kyu Kim, Yonggu Lee, Young-Hyo Lim, Hyun-Kyung Park, Sung Ho Cho, and Seok Hyun Cho. 2020. Non-contact diagnosis of obstructive sleep apnea using impulse-radio ultra-wideband radar. Scientific Reports 10, 1 (2020), 5261. https://doi.org/10.1038/s41598-020-62061-4
[23] Sun Kang、Dong-Kyu Kim、Yonggu Lee、Young-Hyo Lim、Hyun-Kyung Park、Sung Ho Cho 和 Seok Hyun Cho。2020.使用脉冲无线电超宽带雷达非接触式诊断阻塞性睡眠呼吸暂停。科学报告 10, 1 (2020), 5261. https://doi.org/10.1038/s41598-020-62061-4

[24] B. Amir H. Kargar and Mohammad H. Mahoor. 2017. A pilot study on the eBear socially assistive robot: Implication for interacting with elderly people with moderate depression. In 2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids). 756-762. https://doi.org/10.1109/HUMANOIDS. 2017.8246957
[24] B. Amir H. Kargar and Mohammad H. Mahoor.2017.关于 eBear 社交辅助机器人的试点研究：与中度抑郁症老人互动的意义》。2017 IEEE-RAS 第 17 届仿人机器人（Humanoids）国际会议。756-762. https://doi.org/10.1109/HUMANOIDS.2017.8246957

[25] Taeyoung Kil. 2021. Effects of the multimodal intervention program including animal-assisted therapy on depression and self-esteem among university students. Journal of Animal Science and Technology 63, 6 (2021), 1443.
[25] Taeyoung Kil.2021.包括动物辅助治疗在内的多模式干预方案对大学生抑郁和自尊的影响。动物科学与技术杂志》63，6（2021），1443。

[26] Maria Komariah, Shakira Amirah, Emir Gibraltar Faisal, Stephanie Amabella Prayogo, Sidik Maulana, Hesti Platini, Suryani Suryani, Iyus Yosep, and Hidayat Arifin. 2022. Efficacy of internet-based cognitive behavioral therapy for depression and anxiety among global population during the covid-19 pandemic: a systematic review and meta-analysis of a randomized controlled trial study. Healthcare 10, 7 (2022), 1224.
[26] Maria Komariah, Shakira Amirah, Emir Gibraltar Faisal, Stephanie Amabella Prayogo, Sidik Maulana, Hesti Platini, Suryani Suryani, Iyus Yosep, and Hidayat Arifin.2022.基于互联网的认知行为疗法对19型流感大流行期间全球人口抑郁和焦虑的疗效：随机对照试验研究的系统回顾和荟萃分析》。医疗保健》10，7 (2022)，1224。

[27] Georgia Konstantopoulou and Eleni Mavroeidi. 2023. Efficacy of Digital Cognitive Behavior Therapy (DCBT) for the Treatment for Symptoms of Depression. European Journal of Science, Innovation and Technology 3, 3 (2023), 113-119.
[27] Georgia Konstantopoulou and Eleni Mavroeidi.2023.数字认知行为疗法（DCBT）治疗抑郁症状的疗效》。欧洲科学、创新与技术期刊》3，3（2023），113-119。

[28] Kurt Kroenke, Robert L. Spitzer, and Janet BW Williams. 2001. The PHQ-9: validity of a brief depression severity measure. Journal of general internal medicine 16, 9 (2001), 606-613.
[28] Kurt Kroenke、Robert L. Spitzer 和 Janet BW Williams。2001.PHQ-9：简短抑郁严重程度测量的有效性》。普通内科杂志》16，9（2001），606-613。

[29] Timo Lauteslager, Stylianos Kampakis, Adrian J. Williams, Michal Maslik, and Fares Siddiqui. 2020. Performance Evaluation of the Circadia Contactless Breathing Monitor and Sleep Analysis Algorithm for Sleep Stage Classification. In 2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC). IEEE, Montreal, Canada, 5150-5153.
[29] Timo Lauteslager、Stylianos Kampakis、Adrian J. Williams、Michal Maslik 和 Fares Siddiqui。2020.Circadia 非接触式呼吸监测器的性能评估和睡眠阶段分类的睡眠分析算法。2020年第42届电气和电子工程师学会医学与生物学工程学会（EMBC）国际年会。IEEE，加拿大蒙特利尔，5150-5153。

[30] Rachael Luck. 2018. What is it that makes participation in design participatory design? Design studies 59 (2018), 1-8.
[30] Rachael Luck.2018.是什么让参与设计成为参与式设计？Design studies 59 (2018)，1-8.

[31] Mabu Robot Video 2017. The Mabu Personal Healthcare Companion from Catlia Health. Retrieved September 28, 2023 from https://www.youtube.com/watch?v= A3XwzlvOW7k&ab_channel=Welcome.AI
[31] 2017年Mabu机器人视频。Mabu Personal Healthcare Companion from Catlia Health.2023年9月28日检索自https://www.youtube.com/watch?v= A3XwzlvOW7k&ab_channel=Welcome.AI

[32] Dawn A. Marcus. 2013. The science behind animal-assisted therapy. Current pain and headache reports 17 (2013), 1-7.
[32] Dawn A. Marcus.2013.动物辅助疗法背后的科学。当前疼痛与头痛报告 17 (2013)，1-7.

[33] Tyler J. Noorbergen, Marc TP Adam, Mark Roxburgh, and Timm Teubner. 2021. review. AIS Transactions on Human-Computer Interaction 13, 2 (2021), 175-205.
[33] Tyler J. Noorbergen、Marc TP Adam、Mark Roxburgh 和 Timm Teubner。2021. Review.AIS Transactions on Human-Computer Interaction 13, 2 (2021), 175-205.

[34] Farzan M. Noori, Md. Zia Uddin, and Jim Torresen. 2021. Ultra-Wideband Radar-Based Activity Recognition Using Deep Learning. IEEE Access 9 (2021), 138132- 138143. https://doi.org/10.1109/ACCESS.2021.3117667
[34] Farzan M. Noori、Md.Zia Uddin, and Jim Torresen.2021.基于超宽带雷达的深度学习活动识别。IEEE Access 9 (2021), 138132- 138143. https://doi.org/10.1109/ACCESS.2021.3117667

[35] US Department of Health and Human Services. 2023. Major Depression. National Institute of Mental Health (July 2023). https://www.nimh.nih.gov/health/statistics/ major-depression
[35] 美国卫生与公众服务部。2023.重度抑郁症。美国国家精神卫生研究所（2023 年 7 月）。https://www.nimh.nih.gov/health/statistics/ major-depression.

[36] Paro Robot Video 2017. Robot Baby Seal ’Paro’ Helps Elderly Dementia Patients in Hong Kong. Retrieved September 28, 2023 from https://www.youtube.com/ watch?v=g9wj5v22KsA&ab_channel=DailyMail
[36] 2017年帕罗机器人视频。机器人海豹宝宝 "Paro "帮助香港老年痴呆症患者。2023年9月28日取自https://www.youtube.com/ watch?v=g9wj5v22KsA&ab_channel=DailyMail

[37] Dimitris Pnevmatikos, Panagiota Christodoulou, and Nikolaos Fachantidis. 2022. Designing a socially assistive robot for education through a participatory design approach: Pivotal principles for the developers. International Journal of Social Robotics, 14, 3 (2022), 763-788.
[37] Dimitris Pnevmatikos、Panagiota Christodoulou 和 Nikolaos Fachantidis。2022.通过参与式设计方法设计教育社交辅助机器人：开发人员的关键原则。国际社会机器人学杂志》，14，3 (2022)，763-788。

[38] Natasha Randall, Casey C. Bennett, Selma Šabanović, Shinichi Nagata, Lori Eldridge, Sawyer Collins, and Jennifer A. Piatt. 2019. More than just friends: in-home use and design recommendations for sensing socially assistive robots (SARs) by older adults with depression. Paladyn, Journal of Behavioral Robotics, .
[38] Natasha Randall, Casey C. Bennett, Selma Šabanović, Shinichi Nagata, Lori Eldridge, Sawyer Collins, and Jennifer A. Piatt.2019.不仅仅是朋友：患有抑郁症的老年人在家中使用感应社交辅助机器人（SARs）的情况和设计建议》（More than just friends: in-home use and design recommendations for sensing socially assistive robots (SARs) by older adults with depression.Paladyn, Journal of Behavioral Robotics, .

[39] Silvia Rossi, Silvano Junior Santini, Daniela Di Genova, Gianpaolo Maggi, Alberto Verrotti, Giovanni Farello, Roberta Romualdi, Anna Alisi, Alberto Eugenio Tozzi, and Clara Balsano. 2022. Using the social robot NAO for emotional support to children at a pediatric emergency department: Randomized clinical trial. Journal of Medical Internet Research 24, 1 (2022).
[39] Silvia Rossi、Silvano Junior Santini、Daniela Di Genova、Gianpaolo Maggi、Alberto Verrotti、Giovanni Farello、Roberta Romualdi、Anna Alisi、Alberto Eugenio Tozzi 和 Clara Balsano。2022.在儿科急诊室使用社交机器人NAO为儿童提供情感支持：随机临床试验。医学互联网研究杂志》24，1 (2022).

[40] Robert A. Schoevers, Aartjan TF Beekman, D. J. H. Deeg, Cees Jonker, and W. van Tilburg. 2003. Comorbidity and risk-patterns of depression, generalised anxiety disorder and mixed anxiety-depression in later life: results from the AMSTEL study. International journal of geriatric psychiatry 18, 11 (2003), 994-1001.
[40] Robert A. Schoevers, Aartjan TF Beekman, D. J. H. Deeg, Cees Jonker, and W. van Tilburg.2003.晚年抑郁症、广泛性焦虑症和混合型焦虑抑郁症的共存性和风险模式：AMSTEL 研究的结果。International journal of geriatric psychiatry 18, 11 (2003), 994-1001.

[41] Selma, Casey C. Bennett, Wan-Ling Chang, and Lesa Huber. 2013. PARO robot affects diverse interaction modalities in group sensory therapy for older adults with dementia. 2013 IEEE 13th international conference on rehabilitation robotics
[41] Selma、Casey C. Bennett、Wan-Ling Chang 和 Lesa Huber。2013.PARO机器人在老年痴呆症患者的集体感官治疗中影响多种互动模式。2013 IEEE 第 13 届康复机器人国际会议。

(ICORR) (June 2013), 1-6.
(ICORR）（2013 年 6 月），1-6。

[42] Syamimi Shamsuddin, Hanafiah Yussof, Luthfi Idzhar Ismail, Salina Mohamed, Fazah Akhtar Hanapiah, and Nur Ismarrubie Zahari. 2012. Humanoid robot NAO interacting with autistic children of moderately impaired intelligence to augment communication skills. Procedia Engineering 41 (2012), 1533-1538.
[42] Syamimi Shamsuddin, Hanafiah Yussof, Luthfi Idzhar Ismail, Salina Mohamed, Fazah Akhtar Hanapiah, and Nur Ismarrubie Zahari.2012.仿人机器人NAO与智力中度受损的自闭症儿童互动，增强交流技能。程序工程学》41（2012），1533-1538。

[43] Donna M Sudak. 2012. Cognitive behavioral therapy for depression. Psychiatric Clinics 35, 1 (2012), 99-110.
[43] Donna M Sudak.2012.抑郁症的认知行为疗法》。精神病学临床》，35，1（2012），99-110。

[44] Kazue Takayanagi, Takahiro Kirita, and Takanori Shibata. 2014. Comparison of verbal and emotional responses of elderly people with mild/moderate dementia and those with severe dementia in responses to seal robot, PARO. Frontiers in Aging Neuroscience 6 (2014), 247.
[44] Kazue Takayanagi、Takahiro Kirita 和 Takanori Shibata。2014.轻度/中度痴呆症老人与重度痴呆症老人对海豹机器人PARO的语言和情绪反应比较。老龄神经科学前沿 6 (2014)，247。

[45] Katarzyna Tomaszewska, Iga Bomert, and Elżbieta Wilkiewicz-Wawro. 2017. Feline-assisted therapy: Integrating contact with cats into treatment plans. Polish Annals of Medicine 24, 2 (2017), 283-286.
[45] Katarzyna Tomaszewska, Iga Bomert, and Elżbieta Wilkiewicz-Wawro.2017.猫辅助治疗：将与猫的接触纳入治疗计划。波兰医学年鉴》24，2 (2017)，283-286。

[46] World Health Organization WHO. 2023. Depressive disorder. (March 2023). https://www.who.int/news-room/fact- sheets/detail/depression
[46] 世界卫生组织。2023.抑郁症。(https://www.who.int/news-room/fact- sheets/detail/depression.

[47] Renwen Zhang, Jennifer Nicholas, Ashley A Knapp, Andrea K Graham, Elizabeth Gray, Mary J Kwasny, Madhu Reddy, and David C Mohr. 2019. Clinically Meaningful Use of Mental Health Apps and its Effects on Depression: Mixed Methods Study. J Med Internet Res 21, 12 (20 Dec 2019), e15644. https: //doi.org/10.2196/15644
[47] Renwen Zhang, Jennifer Nicholas, Ashley A Knapp, Andrea K Graham, Elizabeth Gray, Mary J Kwasny, Madhu Reddy, and David C Mohr.2019.心理健康应用程序的临床意义及其对抑郁症的影响》：混合方法研究》。J Med Internet Res 21, 12 (20 Dec 2019), e15644. https: //doi.org/10.2196/15644

[48] Selma Šabanović, Wan-Ling Changand Casey C. Bennett, Jennifer A. Piatt, and David Hakken. 2015. A robot of my own: participatory design of socially assistive robots for independently living older adults diagnosed with depression. Human Aspects of IT for the Aged Population. Design for Aging: First International Conference, ITAP 2015, Held as Part of HCI International 2015, Los Angeles, CA, USA, August 2-7, 2015. Proceedings, Part I 1 (2015), 104-114.
[48] Selma Šabanović, Wan-Ling Changand Casey C. Bennett, Jennifer A. Piatt, and David Hakken.2015.我自己的机器人：为被诊断患有抑郁症的独立生活的老年人参与性设计社交辅助机器人》（A robot of my own: participatory design of socially assistive robots for independently living older adults diagnosed with depression.面向老年人群的信息技术的人性化方面》。老龄化设计：第一届国际会议，ITAP 2015，作为2015年国际人机交互大会的一部分举行，美国加利福尼亚州洛杉矶，2015年8月2-7日。论文集，第一部分 1 (2015)，104-114。