The Naughty Drone: A Qualitative Research on Drone as Companion Device
淘气的无人机无人机作为伴侣设备的定性研究

H.5.2 [User Interfaces]: Interaction styles, Style guides, and User-centered design
H.5.2 [用户界面]：交互风格、风格指南和以用户为中心的设计

Design and Human Factors 设计与人为因素

Drone, Companion Technology, Companionship, and Design
无人机、陪伴技术、陪伴和设计

Recently, drones have become the focus of the increasing, worldwide interest. The term "drone" refers to "an aircraft which can be piloted without a human pilot aboard, either remotely or autonomously [16]." Drones, in particular, have wide range of uses and are highly anticipated to have a potential which might
最近，无人机成为全世界日益关注的焦点。所谓 "无人机"，是指 "可以在无人驾驶的情况下，遥控或自主驾驶的飞行器[16]"。无人机尤其具有广泛的用途，并被寄予厚望，其潜力可能包括

allow the usage rate of 2 drones per household to become true [20]. Chris Anderson referred to drone as "a way to increase productivity without adding more people [17]". Drones are generating revenues in the relevant markets such as Amazon, and Google [21, 28, 35, 40]. Drone is increasingly becoming popular as its sizes get smaller, as technology develops, and as the price range becomes diversified [15, 24]. Therefore, as mentioned above, many anticipation and positive predictions on drones are present. However, some cast doubts on the prospect of drone commercialization. Privacy issues and security issues such as 'drone hacking' have appeared, and the safety issue also came up. While such drone-related accidents are causing anxiety among the public and thus discouraging the use of drones, technical progress as well as new regulations for drones are solving security and safety problems to some extent . However, there still remain some parts which are not solved by legal restriction or technological development. Consequently, efforts to get rid of people's anxiety and psychological resistance are necessary to encourage people to use drones [22, 29, 31, 38]. According to Rook, negative experiences usually have greater influences than positive experiences because of their rarer and therefore more salient nature [31]. Consequently, concerns of psychologica resistance on drones must necessarily be answered.
让每户 2 架无人机的使用率成为现实[20]。克里斯-安德森（Chris Anderson）将无人机称为 "在不增加人员的情况下提高生产力的一种方式[17]"。无人机正在为亚马逊和谷歌等相关市场创造收入[21, 28, 35, 40]。随着无人机尺寸的缩小、技术的发展和价格范围的多样化，无人机正变得越来越受欢迎[15, 24]。因此，如上所述，人们对无人机有许多期待和积极的预测。然而，也有人对无人机商业化的前景表示怀疑。隐私问题和 "无人机黑客 "等安全问题已经出现，安全问题也随之而来。虽然这些与无人机相关的事故引起了公众的焦虑，从而阻碍了无人机的使用，但技术进步和新的无人机法规在一定程度上解决了安全和安保问题 。然而，仍有一些问题是法律限制或技术发展无法解决的。因此，要鼓励人们使用无人机，就必须努力消除人们的焦虑和心理抵触[22, 29, 31, 38]。鲁克认为，负面经验通常比正面经验具有更大的影响力，因为负面经验更罕见，因此也更突出[31]。因此，必须回答人们对无人机的心理抵触问题。

These kinds of psychological problems will be solved by adopting companion technology, which is taking center stage as the technology supports emotional well-being and persistent use of the user [2]. While a variety of definitions of the term 'companion" have been suggested, in this paper, a companion refers to an artificial device that spends time with the user, shares some activities, and keeps friendly relation with the user [5, 23 32]. Companionship is a relationship with mutual care and trust, and a pleasing relationship which is incrementally formed and maintained over continuous interactions with a companion , 9]. According to Nass, findings that the robots can be recognized as social actors and interact with people, underpin the fact that robots or intelligent device can sufficiently perform a role of social companion [27, 30]. Additionally, a number of robots which were designed to engage in social human interaction have been released [39], and a great majority of people were fond of the idea of a robot companion [7]. As Benyon referred, if the user built the companionship with the artificial companion through companion technology, trust would be planned, perceived risk toward the counterpart would be reduced, and the emotional wellbeing of the user would be positively affected [2]. Consequently, the companionship will alleviate the psychological resistance toward the object and have a beneficial effect on the persistent use of the object in the end. Therefore, in order to settle the problem regarding psychological resistance, the drone has to be designed as an artificial companion.
这些心理问题将通过采用伴侣技术得到解决，因为该技术有助于用户的情感健康和持续使用[2]。虽然人们对 "伴侣 "一词提出了多种定义，但在本文中，伴侣指的是与用户共度时光、分享某些活动并与用户保持友好关系的人工设备[5, 23 32]。陪伴是一种相互关心、相互信任的关系，是在与同伴的持续互动中逐步形成并保持的一种愉悦关系 ，9]。纳斯认为，机器人可以被识别为社会行动者并与人互动的研究结果证明，机器人或智能设备可以充分扮演社会陪伴者的角色[27, 30]。此外，一些旨在参与人类社交互动的机器人已经面世[39]，绝大多数人都喜欢机器人伴侣的概念[7]。正如 Benyon 所说，如果用户通过伴侣技术与人工伴侣建立起伴侣关系，信任就会得到规划，对对方的感知风险就会降低，用户的情绪健康也会受到积极影响[2]。因此，陪伴会减轻用户对人工对象的心理抵触情绪，并对用户最终持续使用人工对象产生有利影响。因此，为了解决心理抵触问题，必须将无人机设计成人工伴侣。

Besides, the possibilities for new elements of life companion to emerge from drones are strong on account of differentiated characteristics of drones which are unlike traditional devices, such as flying free, possessing certain level of autonomy and emotional factors, offering instrumental support, and having the potential of downsizing. Thus, if we can utilize the above nature of drones, drones will perform better as companion than other current companion robots. In other words, so as to resolve the psychological issues of drones and to achieve sustainable use of drones, drones should play a role as companion, and we believe that drones are very suitable for companion device.
此外，由于无人机具有不同于传统设备的差异化特征，如自由飞行、具有一定的自主性和情感因素、提供工具支持、具有缩小体积的潜力等，无人机出现新的生活伴侣元素的可能性很大。因此，如果我们能利用无人机的上述特性，无人机作为伴侣的表现将优于目前的其他伴侣机器人。换句话说，为了解决无人机的心理问题，实现无人机的可持续使用，无人机应该发挥陪伴的作用，我们认为无人机非常适合作为陪伴设备。

Even though there has been a great deal of attention on drones recently, much of the studies on drones have focused on technical parts , and still have not treated drones as companion in much detail [25]. Therefore, further research on how the user feels when interacting with a drone and how the emotion ultimately influences companionship is needed, because the emotional factors, such as mutual caring and trust, are crucial part of building companionship [13]. Thus, we tried to identify what characteristics people desire for drones to perceive them as companion.
尽管最近对无人机的关注度很高，但大部分关于无人机的研究都集中在技术部分 ，仍然没有将无人机作为伴侣进行详细研究[25]。因此，需要进一步研究用户在与无人机互动时的感受，以及情感如何最终影响陪伴关系，因为情感因素（如相互关心和信任）是建立陪伴关系的关键部分[13]。因此，我们试图找出人们希望无人机具备哪些特征，从而将其视为同伴。

The aims of this research have, therefore, been to develop the understanding of perception on drones, and to investigate desired characteristics for drones as artificial companion, and as life companion device by extension. We adopted a human-centered approach to investigate people's perceptions and desires of companion device, and explored the following research question: what characteristics drones as artificial companion should have?
因此，本研究的目的是了解人们对无人机的看法，并调查无人机作为人工伴侣以及作为生活伴侣设备的理想特性。我们采用以人为本的方法来调查人们对伴侣设备的看法和愿望，并探讨以下研究问题：作为人工伴侣的无人机应具备哪些特征？

We have investigated our research from a qualitative perspective for several reasons [6]. Firstly, we explored areas which are not thoroughly researched yet, since drone area is an unexplored field and is still realized as novel technology by people. Secondly, we wanted to explore the inner experiences of participants and figure out how meanings about drones are formed and transformed. To sum up, we sought to examine the perception and the understanding of drone which is new and unexplored area by using qualitative research. By doing this, we tried to draw both user experience factors on drones and to suggest principles that may help people recognize drones as companion device.
我们从定性的角度进行研究有几个原因[6]。首先，我们探索了尚未深入研究的领域，因为无人机领域是一个尚未开发的领域，人们仍将其视为一项新技术。其次，我们希望探索参与者的内心体验，弄清关于无人机的意义是如何形成和转变的。总之，我们试图通过定性研究来考察人们对无人机这一尚未开发的新领域的看法和理解。通过这样做，我们试图得出无人机的用户体验因素，并提出有助于人们认识无人机作为伴侣设备的原则。

10 participants completed the study. The study was conducted in the form of unstructured in-depth interview. First three participants were recruited by personal network, and the rest seven participants were recruited by snow sampling from the first three. After conducting interviews of 7 people, 3 people were recruited again to check if no new codes or themes arose. We interviewed five men and five women, with an age range of 23-28 years ( ). Questions were open-ended, and every behavior of participants was video-recorded in order to focus on the interaction between participants and drones, and subsequently transcribed. 1) Participants were given a piece of paper, and received instruction to draw drones in their mind and to write their feelings on drones, or additional features that he or she wants to add freely. This was performed in intention to understand how people usually think about drones. 2) We gave participants a chance to manipulate the drone personally in order to provide actual experience for who did not have experience of drones, and to form rapport with drone, then we observed the interaction between participants and the drone. 3) Then, participants were instructed to think as if drone is automatically flying by itself when the drone was operated by one of our researcher. 4) During the drone was controlled automatically, we gave participants a task of naming drone to check up on how companionship on drones can be made or shown, and we examined the reactions of participants as the drone approached to the participant when their name was called by the participant. 5) During the whole process, participants were asked to say everything he or she had in mind out loud. 6) Then, the words and reactions acquired from previous steps were inquired for more specific reasons, and we strived to find out which elements are needed to help people regard drone as companion. 7) Finally, participants were given the paper that he or she had been given at the first phase of the interview, and were instructed to modify the drawings or writings if they wanted to. The interviews were concluded after the wrap-up questions. Participants were invited to our research and the first phase of interviews was conducted inside. The middle phase, at which drones were actually operated, was conducted outside in order to consider real interactive situation with drones. The interviews lasted for 40 minutes on average, and the interviewees were compensated for their participation (as worth as ).
10 名参与者完成了研究。研究以非结构化深度访谈的形式进行。前 3 位参与者是通过个人网络招募的，其余 7 位参与者是从前 3 位参与者中通过雪地抽样招募的。在对 7 人进行访谈后，我们再次招募了 3 人，以检查是否出现了新的代码或主题。我们采访了 5 男 5 女，年龄在 23-28 岁之间（ ）。问题是开放式的，参与者的每一个行为都被录制下来，以便关注参与者与无人机之间的互动，随后进行转录。1) 给参与者一张纸，指导他们在脑海中描绘无人机，并写下自己对无人机的感受或想自由添加的功能。这样做的目的是了解人们通常是如何看待无人机的。2) 为了让没有无人机使用经验的人有实际体验，并与无人机形成默契，我们给了参与者亲自操控无人机的机会，然后观察参与者与无人机之间的互动。3) 然后，由我们的研究人员操作无人机，让参与者认为无人机是自动飞行的。4) 在无人机被自动控制的过程中，我们给参与者布置了一个为无人机命名的任务，以考察无人机上的陪伴关系是如何产生或表现出来的。5) 在整个过程中，我们要求参与者大声说出自己心中所想的一切。 6) 然后，我们就前几个步骤中获得的言语和反应询问了更具体的原因，并努力找出帮助人们将无人机视为同伴所需的要素。7) 最后，我们将第一阶段的访谈纸发给参与者，并指导他们根据自己的意愿修改图画或文字。在总结问题之后，访谈结束。参与者受邀参加我们的研究，第一阶段的访谈在室内进行。中间阶段是实际操作无人机，在室外进行，以考虑与无人机的真实互动情况。访谈平均持续 40 分钟，受访者获得了参与报酬（价值为 ）。

Additionally, for those three interviewees who participated in additional interview, we used images to resonate (trigger) and to identify the concept in much more depth. 28 images of characters and 87 cards of personality [36] were given and participants were asked to check the cards and personalities which he or she might think that drone would possess. As for the images of characters, characters which came up in conversation from previous interviews were chosen, and the characters which appeared in the movies were also added, because the concept of companion was frequently shown in the movies or literature [2]. In order to analyze the data obtained from the above process, we used open coding, axial coding and selective coding analysis on the transcripts.
此外，对于参与额外访谈的三位受访者，我们使用了图片来引起共鸣（触发），并更深入地确定概念。我们给出了 28 张人物形象图片和 87 张个性卡片[36]，要求受访者勾选自己认为无人机可能拥有的卡片和个性。至于人物形象，我们选择了以前访谈中出现过的人物形象，还加入了电影中出现过的人物形象，因为 "同伴 "这一概念经常出现在电影或文学作品中[2]。为了对上述过程中获得的数据进行分析，我们对访谈记录采用了开放式编码、轴向编码和选择性编码分析。

Figure 1. Example of Study design: drawing, showing drone, controlling drone in person, and automatically
图 1.研究设计示例：绘图、展示无人机、亲自控制无人机和自动控制无人机

We gave an overview of the data, discussed several themes that emerged during analysis, and related the findings to our research questions. Generally, we presented our results by the functionality, provision of social role, and attribution of personality.
我们概述了数据，讨论了分析过程中出现的几个主题，并将结果与我们的研究问题联系起来。一般来说，我们按照功能、社会角色的提供和个性归因来介绍我们的结果。

Most participants replied that they will likely to use drone when they perceive value from it. For utilitarian value, participants
大多数与会者回答说，当他们认为无人机有价值时，他们就有可能使用无人机。对于功利价值，参与者
hoped to receive practical support from drones such as taking necessary information and giving errands at a suitable time and place. On the other hand, for hedonic value, people pursued fun and interests to be satisfied by playing game using drones. Furthermore, there were some cases that utilitarian value and hedonic value were pursued at the same time, such as video shooting through drones and that are linked to one's smartphone.
希望得到无人机的实际支持，如在合适的时间和地点获取必要的信息和提供服务。另一方面，在享乐价值方面，人们通过使用无人机玩游戏来追求乐趣和兴趣的满足。此外，还有一些同时追求功利价值和享乐价值的情况，如通过无人机拍摄视频并与智能手机相连。

Functions focused on the utilitarian value generally utilize the unique properties of drone's flying, or filming function. Various utilitarian functions were proposed, and one of them was the transportational service, which includes delivery service, doing errands (Drone-through) and navigating service, utilizing the drone's feature of flying free. There were also other functions including real time monitoring and surveillance function, utilizing the unique properties of drone's flying and filming function. We named this function "the third eye".
注重功利价值的功能一般利用无人机飞行的独特性能或拍摄功能。人们提出了各种实用功能，其中之一是运输服务，包括送货服务、跑腿（无人机穿越）和导航服务，利用了无人机自由飞行的特点。还有其他一些功能，包括利用无人机飞行和拍摄功能的独特特性实现实时监控和监视功能。我们将这一功能命名为 "第三只眼"。

P1: [After controlling drone] I guess current drones are limited to the function of photographing, but I want other functions to be enabled. For example, music can be played by the drone, or I can make him buy coffee and deliver it to me. (Laugh) Maybe, there will be markets only for drones. Drone-through! It would be fun!
P1：[控制无人机后]我想现在的无人机仅限于拍照功能，但我希望还能实现其他功能。比如，无人机可以播放音乐，或者我可以让它买咖啡，然后给我送过来。(笑）也许，以后会有只属于无人机的市场。Drone-through！那会很有趣！

Functions focused on hedonic value include the case of singing drones, participating in drone racing, and being collected by kidult. In the case of "the third eye", the function satisfies both hedonic and utilitarian value, thanks to the camera built in the drones. In addition, there was a response from one of the participants that she wants a drone which not only works as an air freshener, but also welcomes the participant when she gets home (P5). Functions of drones that are currently on the market are usually focused on the one of the two; utilitarian, or hedonic value And these two functions are pretty limited. However, future drones need to extend both value to some extent. As the response below speaks, people wanted the drone which has functions that fill up both utilitarian and hedonic value, when they were asked about their desired attributes of drones using characters by analogy.
注重享乐价值的功能包括无人机唱歌、参加无人机比赛和被 kidult 收集。在 "第三只眼 "的案例中，由于无人机内置了摄像头，该功能同时满足了享乐价值和功利价值。此外，一位参与者还表示，她希望无人机不仅能用作空气清新剂，还能在她回家时欢迎她（P5）。目前市场上的无人机功能通常集中在功利或享乐价值中的一种，而这两种功能是相当有限的。然而，未来的无人机需要在一定程度上扩展这两种价值。正如下面的回答所表明的那样，在通过类比人物来询问人们对无人机属性的期望时，人们希望无人机能够同时满足功利和享乐两种价值。

P10: [Shown characters] I think the drone can be compared to Tars from Interstellar, or Jarvis from Iron Man. Smart and helpful? Um, in case of Tars, I can manipulate the level of humor, and apply the characteristics of puppy or Minions. With those traits, it would be nice to get necessary information immediately.
P10: [展示人物]我认为无人机可以比作《星际穿越》中的塔斯，或者《钢铁侠》中的贾维斯。聪明又乐于助人？嗯，如果是塔斯，我可以控制幽默的程度，并应用小狗或奴才的特性。有了这些特性，就可以立即获得必要的信息了。

According to Benyon, functions of artificial companions available in commercial market tend to belong to one of four categories: entertainment, educational, informative, or medical [2]. However, the interviews on drones revealed that people want drones which perform a number of categories of function instead of only one. Consequently, in order to form reciprocal relationship, and forge closer ties, people should be able to perceive value from the object [3]. Therefore functions, which can make people perceive value from the object, will play a significant role in having companionship with the object.
根据 Benyon 的观点，商业市场上的人工伴侣功能往往属于四类之一：娱乐、教育、信息或医疗[2]。然而，关于无人机的访谈显示，人们希望无人机能实现多种功能，而不是只有一种。因此，为了形成互惠关系，建立更紧密的联系，人们应该能够从物品中感知到价值[3]。因此，能让人们从物体中感知价值的功能将在与物体建立伙伴关系方面发挥重要作用。

If people can consider the object as a social intelligent being, people will be able to provide a role or personality to the object [18]. We asked participants to say feelings or characteristics of drones freely, and we could divide appeared characteristics into 4 groups depending on the level of perceived controllability. This will enable us to grant the role which represents each group of applicable characteristics. We asked 3 participants from additional interview more specifically about the roles that would be the best suit for the drone and attempted to confirm whether roles properly represent the group of characteristics. The roles can be classified as servant, pet, friend, and bully, in order of high level of perceived controllability. Firstly, the bully, of which the level of controllability is the lowest, represents responses such as danger, fear, anxiety, hostility, and excessive unpredictability on drones. When faced the drone for the first time, one of the participants showed hostility toward the drone and could not regard the drone as a companion. Likewise, if people perceive the object as a bully, it would be hard to feel companionship with the object, so we determined that the role of bully will not belong within the boundary of companion.
如果人们能将物体视为一种社会智能存在，那么人们就能为物体赋予角色或个性[18]。我们要求参与者自由说出对无人机的感受或特征，并根据可控程度将出现的特征分为 4 组。这样，我们就能赋予代表每组适用特征的角色。我们从额外的访谈中向 3 位参与者详细询问了最适合无人机的角色，并试图确认这些角色是否恰当地代表了各组特征。按照可控性的高低，这些角色可分为仆人、宠物、朋友和恶霸。首先，可控性水平最低的 "恶霸 "代表了对无人机的危险、恐惧、焦虑、敌意和过度不可预测性等反应。当第一次面对无人机时，一位参与者对无人机表现出敌意，无法将无人机视为同伴。同样，如果人们认为无人机是一个欺凌者，那么就很难感受到与无人机之间的同伴关系，因此我们认为欺凌者的角色不属于同伴的范畴。

P6: [During watching the drone being controlled automatically] Even though I know that I will not be hit by that, I feel like that he is going to attack me. (When the drone rotates and looks at the participant) Don't look at me! Staring at me is even worse than before. Feel like I am the target. He looks so bad.
P6：[在观看无人机自动控制的过程中]虽然我知道我不会被它击中，但我感觉它要攻击我。(当无人机旋转并看着参与者时）别看我！盯着我看比以前更难受。感觉自己成了靶子。他看起来很糟糕。

Secondly, servant role, of which perceived controllability is the highest, represents responses such as reliable, obedient, trust, and faithful. When people want drones to play a role of servant, they want to be offered proper function, rather than conversation or interaction with drones.
其次，仆人角色，其中感知可控性最高，代表了可靠、服从、信任和忠诚等反应。当人们希望无人机扮演仆人角色时，他们希望得到的是适当的功能，而不是与无人机对话或互动。

P9: [Shown characters] I have a feeling that something keeps following me... And cute, um, but he seems he is able to work. Although he will not work much, he can be a slave, and I can put him to do something that I cannot.
P9: [显示字符] 我感觉有什么东西一直跟着我......而且很可爱，嗯，不过他好像还能干活。虽然他工作不多，但他可以做奴隶，我可以让他做我做不了的事。

Thirdly, pet, with the next higher perceived controllability, represents the responses such as dependent, adorable, worrying, minding, in need for others' care, somewhat bothering, weak, and a little unpredictability.
第三，宠物的可控性次之，代表了依赖、可爱、担心、介意、需要他人照顾、有点麻烦、弱小和有点不可预测等反应。

P5: [Wrap-up question] I hope him to be safe. I hope him to control himself to be safe. I don't want him to die... Animals sense danger, but he doesn't... If I have to fix him when he is broken, I would feel really bad.
P5: [总结问题] 我希望他平安无事。我希望他能控制自己，保证安全。我不希望他死...动物能感觉到危险，但他感觉不到...如果我必须在它坏掉的时候把它修好，我会很难过。

P9: [Shown characters] I want my drone to do something interesting. I do not mind if he doesn't work much. You know, servant must work professionally as I told him to do. And the drone is in lower class. That's why he is not my friend. By the way, he has to listen to my order, so he is not in the bully group. Therefore I want him to act like pet.
P9：[显示字符] 我希望我的无人机能做些有趣的事情。我不介意他工作不多。你知道，仆人必须按照我的要求专业地工作。而无人机是下等人。所以他不是我的朋友。顺便说一句，他必须听我的命令，所以他不属于恶霸集团。因此，我希望他表现得像宠物一样。

The next role is friend, of which perceived controllability is lower than pet but bigger than bully, represents the responses such as autonomous, same (status), and moderate level of unpredictability.
下一个角色是朋友，其感知可控性低于宠物，但大于恶霸，代表了自主、相同（地位）和中等程度的不可预测性。

P10: My ideas can be limited, because I am a human. But if he (the drone) is knowledgeable, he can be an argumentative fellow and refute my opinion. I like that more... He can correct my thoughts and suggest various possibilities ... Well, I want him to lash back my words, so he is more close to friend... not obedient at all.
P10：我的想法可能有局限性，因为我是人。但如果他（无人机）知识渊博，他可以成为一个善于争辩的家伙，反驳我的观点。我更喜欢这样...他可以纠正我的想法，提出各种可能性......好吧，我希望他能反驳我的话，这样他就更像朋友了......一点也不听话。

The majority of participants did not want drones to function as only one role. Rather, they looked forward to have drones which play various roles situationally. As mentioned in the interviews, people wanted different role of drone depending on situation and contexts. According to Goetz et al., people anticipated robots to look and act appropriately for different tasks [14]. In addition, we first assumed that participants will prefer the drone with the highest controllability, and that the servant role will be most
大多数与会者不希望无人机只扮演一种角色。相反，他们希望无人机能在不同情况下扮演不同的角色。正如访谈中提到的，人们希望无人机能根据不同的情况和背景扮演不同的角色。根据 Goetz 等人的研究，人们希望机器人在执行不同任务时有适当的外观和行为[14]。此外，我们首先假定参与者会更喜欢可控性最高的无人机，而仆人角色将是最受欢迎的。
preferred. However, in contrast to our expectations, there were also many participants who favored the drone which works as a friend or a pet. It was interesting that the users can bear unpredictability of the drone to a certain degree.
他们更喜欢无人机。然而，与我们的预期相反，也有许多参与者对作为朋友或宠物的无人机情有独钟。有趣的是，用户可以在一定程度上承受无人机的不可预测性。

P1: [After controlling drone] Metaphorically speaking, it feels like he (the drone) is following after me in real time... like Pikachu? I want him to have his own function and have a little bit of loyalty. Sometimes he will be equal and sometimes he will be loyal to me. Then, he can be a companion who is also a servant. Wait, but I need affection from him. I want him to be lovable as well as loyal to me.
P1：[控制无人机后]比喻说，感觉他（无人机）在实时跟踪我......就像皮卡丘？我希望他有自己的功能，并且有一点忠诚。有时他是平等的，有时他对我忠诚。这样，他就能成为一个既是伙伴又是仆人的人了。等等，但我需要他的爱。我希望他既可爱又对我忠诚。

P10: [Shown characters] Somewhat between servant and friend? He is like servant when he offers necessary information or gives me a hand when I demand some help. Nevertheless, he can be my friend, especially when I think him as Wall-E. So, servant and friend ... I don't know. It's ambiguous. It's hard to position drone in one of four roles that you suggested.
P10: [显示人物]介于仆人和朋友之间？当他提供必要的信息，或者在我需要帮助的时候帮我一把时，他就像仆人。不过，他也可以是我的朋友，尤其是当我把他当成 Wall-E 的时候。那么，仆人和朋友......我不知道。这很模糊。很难将无人机定位为你建议的四种角色之一。

Similarly, with regard to level of difficulty when controlling the drone personally, most expressed difficulty controlling the drone, and wanted improvements of the drone to move the way they wanted to. However, this doesn't mean that fully automation or unchallenging is wanted. Half of the participants wanted to feel the pleasure of conquest by raising the level of difficulty in control (P1, 2, 4, 7, and 9). This finding may have relevance to giving one's own pet training to obey the owner [1].
同样，关于个人控制无人机的困难程度，大多数人表示控制无人机有困难，希望改进无人机的移动方式。不过，这并不意味着希望实现完全自动化或毫无挑战性。半数参与者希望通过提高控制难度来感受征服的乐趣（P1、2、4、7 和 9）。这一发现可能与训练自己的宠物服从主人有关[1]。

P7: [During controlling personally] (Q: shall we go inside and continue the interview?) Um, can I try controlling the drone one more time, please?
P7：[亲自控制过程中]（问：我们要进去继续采访吗）嗯，我能再试一次控制无人机吗？

As a result, characteristics of drones are extended over all the three roles which come under boundary of companion. That is, owing to the fact that the drone is a convenient device to manipulate the level of perceived controllability, it is not difficult to set the drone to the role of companion that people want. This means that possibility of utilizing drone as companion is high.
因此，无人机的特点被扩展到了属于 "伴侣 "范畴的所有三种角色。也就是说，由于无人机是一种可以方便操控的设备，因此不难将无人机设定为人们想要的同伴角色。这意味着利用无人机作为伴侣的可能性很高。

The role of a single companion device could be changed depending on the situation. However, the personality or trait that people wanted from them showed consistency in some degree. The followings appeared repeatedly during the interview: tolerable and partial predictability, limited controllability, adorability, necessity for others' care, and consistent autonomy. Then, we found out that these characteristics which emerged repeatedly can be clustered and be comprised in a distinct and unique personality. We named this personality "naughty".
单个伴侣设备的角色可以根据情况改变。不过，人们希望从它们身上获得的个性或特质在某种程度上是一致的。在访谈中反复出现的是：可容忍和部分可预测性、有限的可控性、可爱性、需要他人照顾、持续的自主性。然后，我们发现，这些反复出现的特征可以集中起来，形成一种独特的个性。我们将这种性格命名为 "淘气"。

P2: [During watching the drone being controlled automatically] Feeling like a naughty cat (laugh)
P2：[在观看无人机被自动控制的过程中]感觉自己像只淘气的猫（笑）。

P5: [After controlling drone] Naughty boy would be better. It would be a lot easier and more convenient if the drone listens to my word. However, to tell the truth, I don't have much job to give him (the drone). I know he can't clean up my room. I can do a lot from my own, so I think something interesting would be better, and I would rather want the drone to be little naughty.
P5: [控制无人机后] 还是淘气包好。如果无人机听我的话，那就更简单方便了。不过，说实话，我也没什么工作可以给他（无人机）做。我知道他不能打扫我的房间。我自己可以做很多事情，所以我觉得有趣的事情会更好，我更希望无人机能淘气一点。

P10: In his case, when I saw him flying alone, being bumped, falling, and then flying again ... seemed little stupid but cute.
P10: 就他而言，当我看到他独自飞行时，被撞了一下，摔倒了，然后又飞了起来......看起来有点傻，但很可爱。

One of the participants responded that the personality of the drone is distractive and unfocused on account of the differentiated function which enables drones to fly free in all directions (P9).
其中一位与会者回答说，无人机的个性使人分心，注意力不集中，这是因为无人机的差异化功能使其可以自由地向各个方向飞行（P9）。
For those three interviewees from additional interview, we asked people to select characters that resemble the drone, and there were couple of characters which were selected commonly: Minions and Wall-E. The common ground of these two was that they also have features we have mentioned above, such as tolerable unpredictability, adorability, necessity for others' care and consistent autonomy.
对于额外访谈中的三位受访者，我们要求他们选择与无人机相似的角色，其中有几个角色被普遍选中：Minions 和 Wall-E。这两个角色的共同点是，他们也具有我们上文提到的特点，如可以忍受的不可预测性、可爱性、需要他人照顾和一贯的自主性。

P9: [Shown characters] I want him to be Minions. Cute, following me every day... Yes, those kinds of things...
P9: [展示角色] 我想让他成为迷你族。可爱，每天跟着我...是的，就是这类东西......

P10: [Shown characters] I want him to be silly and whimsical such as Minions.
P10：[展示角色]我想让他像迷你族那样傻乎乎、异想天开。

P1: [After controlling drone] I want some parts of him to be naughty. I have tried to control, but it wasn't easy and he could be more accurate.
P1：[控制无人机后]我想让它的某些部分淘气起来。我已经试着控制了，但这并不容易，他还可以更精确一些。

Since the opinion of people who wanted the drone to have personality and traits of naughty came out continuously (21 codes), this should be taken in consideration when drones are designed as a companion. In fact, the personality has been treated very importantly in the computer agent [26]. When the companionship is formed, people will expect to receive what they want from the device which took on the personality [2]. In addition, these findings are in line with previous research on anthropomorphism , which suggests that implementing features of unpredictability in the behavior of robot can make an illusion of it to seem more natural and alive.
由于希望无人机具有个性和淘气特质的意见不断涌现（21 个代码），因此在设计无人机作为伴侣时应考虑到这一点。事实上，在计算机代理中，个性已经得到了非常重要的处理[26]。当伴侣关系形成时，人们会期望从具有人格的设备中得到他们想要的东西[2]。此外，这些研究结果与之前关于拟人化 的研究结果一致，即在机器人的行为中加入不可预测的特征，可以让人产生一种错觉，认为机器人看起来更自然、更有生命力。

Our study has many differentiated features in comparison with others. Firstly, we tried to find what characteristics companion device has, and hence discovered relevant variables. Secondly, we found that people provide different roles depending on the level of perceived controllability. In view of the results that are so far identified about the companion technology, many studies have focused on the companion who plays a role of an assistant or a servant, and study on a companion device as a friend or pet was not dealt well with. Furthermore, according to the previous study, there were few people who want the robot companion act as a friend [7]. However, many people wanted their drone to work as a companion who acts like a friend or a pet, so further study regarding this issue is needed. Next, we made an attempt to research drone, which is an unexplored field and which is struggling with a negative view on commercial side currently. In addition, we also discovered counter intuitive insight that the personality of naughty was preferred for drone when the drone works as companion. This finding conflicts with earlier studies, which revealed that the most people wanted predictable robot companion [7]. However, earlier studies mainly examined the robot companion in situation of an assistant or a butler, not a friend, and we inferred that this affected the results. Finally, we also found that current market has divided the drone market into utilitarian function and hedonic function, and focused on only one of them. However, we could identify that people wanted both utilitarian and hedonic function simultaneously in some degree. During the in-depth interview, we tried to proceed the interview session creatively, such as by making people to control drone personally, and to draw images of drone. We also discovered a new insight through giving a task of granting name to the drone.
与其他研究相比，我们的研究具有许多与众不同的特点。首先，我们试图找出伴侣设备的特点，从而发现相关变量。其次，我们发现人们会根据感知到的可控性水平提供不同的角色。从目前已确定的有关伴侣技术的结果来看，许多研究都集中在扮演助手或仆人角色的伴侣上，而对作为朋友或宠物的伴侣设备的研究并没有得到很好的处理。此外，根据之前的研究，很少有人希望机器人伴侣扮演朋友的角色[7]。然而，很多人希望无人机能像朋友或宠物一样作为伴侣工作，因此需要对这一问题进行进一步研究。接下来，我们尝试对无人机进行研究，这是一个尚未开发的领域，目前在商业方面正与负面观点作斗争。此外，我们还发现了与直觉相反的观点，即当无人机作为伴侣工作时，淘气的个性更受青睐。这一发现与早前的研究相冲突，早前的研究显示，大多数人都希望机器人同伴具有可预测性[7]。不过，先前的研究主要考察了机器人伴侣作为助手或管家的情况，而不是作为朋友的情况，我们推断这影响了研究结果。最后，我们还发现，目前的市场将无人机市场分为功利性功能和享乐性功能，并只关注其中一种。然而，我们可以发现，人们在某种程度上同时希望获得功利性和享乐性功能。 在深度访谈过程中，我们尝试创造性地开展访谈环节，如让人们亲自操控无人机、绘制无人机图像等。我们还通过给无人机取名字的任务发现了新的见解。

Based on the results, we proposed three solutions when designing drone as a companion: 1. The drone should possess functions that fill up both utilitarian and hedonic value. That means, for example,
根据研究结果，我们提出了将无人机作为伴侣设计的三种解决方案：1.无人机应具备充实功利和享乐价值的功能。例如
people want not only the drone which can take a photo, but also the drone which can enjoy and share sports activity with them.
人们不仅想要能拍照的无人机，还想要能与他们一起享受和分享体育活动的无人机。

There are several limitations to the study we have presented. First, our participants were concentrated in the age of twenties due to the snowball sampling. Therefore, the users from the other ages need to be investigated. Second, there might be a bias from the experiment equipment that we used. Further investigation is needed to identify whether the model of the drone had affected the results. Lastly, companion device can cause a significant change in the relationships between people and technologies [2]. Therefore, further research regarding moral and ethical issues should be covered. In the future, we would like to develop the comprehensive understanding of drone as companion, and investigate requisite factors that make people perceive drone as a companion, and hence build the conceptual model from findings.
我们所做的研究有几个局限性。首先，由于采用了 "滚雪球 "式抽样，我们的参与者主要集中在 20 多岁。因此，还需要对其他年龄段的用户进行调查。其次，我们使用的实验设备可能存在偏差。我们需要进一步调查无人机的型号是否影响了实验结果。最后，伴侣设备会导致人与技术之间的关系发生重大变化[2]。因此，应进一步研究道德和伦理问题。未来，我们希望对无人机作为伴侣有一个全面的了解，并调查使人们将无人机视为伴侣的必要因素，从而根据调查结果建立概念模型。

We presented a qualitative research study of perception on drones and characteristics of drone as a companion device. Through a series of 10 interviews, we discussed the reactions to the drone and several themes regarding artificial companion. We found that people wanted to perceive both utilitarian and hedonic value from the drone and people also provide various roles to the drone depending on the situation. Major findings include that most people wanted the drone which has characteristics of tolerable predictability, limited controllability, adorability, necessity for others' care and consistent autonomy. We comprised these traits into a unique personality as "naughty". We recommend considering these findings when designing drone as a companion. Since the theoretical framework in relation to the companion technology does not exist in current studies, we will make an attempt to build the model and examine the requisites for life companion device.
我们介绍了一项关于对无人机的看法以及无人机作为伴侣设备的特点的定性研究。通过 10 个系列访谈，我们讨论了人们对无人机的反应以及有关人工伴侣的几个主题。我们发现，人们希望从无人机中感知功利和享乐价值，人们还根据不同情况赋予无人机不同的角色。主要发现包括：大多数人希望无人机具有可容忍的可预测性、有限的可控性、可爱性、需要他人照顾和持续的自主性等特征。我们将这些特征归纳为 "淘气 "这一独特个性。我们建议在设计作为伴侣的无人机时考虑这些发现。由于目前的研究中还不存在与伴侣技术相关的理论框架，我们将尝试建立模型并研究生活伴侣设备的必要条件。

This research was supported by the MSIP, Korea, under the GITRC support program (IITP-2015-R6812-15-0001) supervised by the IITP.
这项研究得到了韩国 MSIP 的支持，该项目是由韩国工业技术研究院监督的 GITRC 支持项目（IITP-2015-R6812-15-0001）的一部分。

[1] Bennett, P. C., Cooper, N., Rohlf, V. I., & Mornement, K. 2007. Factors influencing owner satisfaction with companion-dog-training facilities. Journal of Applied Animal Welfare Science, 10(3), 217-241. DOI= http://doi.acm.org/10.1080/10888700701353626.
[1] Bennett, P. C., Cooper, N., Rohlf, V. I., & Mornement, K. 2007.影响主人对伴侣犬训练设施满意度的因素。应用动物福利科学杂志》，10（3），217-241。DOI= http://doi.acm.org/10.1080/10888700701353626.

[2] Benyon, D., & Mival, O. 2010. From human-computer interactions to human-companion relationships. In
[2] Benyon, D., & Mival, O. 2010.从人机交互到人伴关系。见
Proceedings of the First International Conference on Intelligent Interactive Technologies and Multimedia (IITM '10), M. D. Tiwari, R. C. Tripathi, and Anupam Agrawal (Eds.). ACM, New York, NY, USA, 1-9. DOI=http://dx.doi.org/10.1145/1963564.1963565.
《第一届智能交互技术与多媒体国际会议（IITM '10）论文集》，M. D. Tiwari、R. C. Tripathi 和 Anupam Agrawal（编辑）。ACM, New York, NY, USA, 1-9.DOI=http://dx.doi.org/10.1145/1963564.1963565.

[3] Bickmore, T. W., & Picard, R. W. 2005. Establishing and maintaining long-term human-computer relationships. Trans. Comput.-Hum. Interact. 12, 2 (June 2005), 293-327. DOI=http://dx.doi.org/10.1145/1067860.1067867.
[3] Bickmore, T. W., & Picard, R. W. 2005.建立并维持长期的人机关系。 Trans.Comput.-Hum.Interact.12, 2 (June 2005), 293-327.DOI=http://dx.doi.org/10.1145/1067860.1067867.

[4] Bulis K. 2015. Hybrid Power Could Help Drone Delivery Take Off. MIT Technology Review. (Mar. 2015), Retrieved from
[4] Bulis K. 2015.混合动力可助无人机送货起飞》。麻省理工学院技术评论》。(Mar. 2015), Retrieved from

[5] Calderon, C. A. A., Norouzi, A., & Qi, Z. 2014. Humanoid social robots: A design for interaction. In Control Automation Robotics & Vision (ICARCV), 2014 13th International Conference on. (Dec. 2014), 871-876. DOI=http://dx.doi.org/10.1109/ICARCV.2014.7064419.
[5] Calderon, C. A. A., Norouzi, A., & Qi, Z. 2014.仿人社交机器人：交互设计。In Control Automation Robotics & Vision (ICARCV), 2014 13th International Conference on.(Dec. 2014), 871-876.DOI=http://dx.doi.org/10.1109/ICARCV.2014.7064419.

[6] Corbin, J., & Strauss, A. 2014. Basics of qualitative research: Techniques and procedures for developing grounded theory. Sage publications.
[6] Corbin, J., & Strauss, A. 2014.Basics of qualitative research：发展基础理论的技术与程序》。Sage publications.

[7] Dautenhahn, K., Woods, S., Kaouri, C., Walters, M., Koay, K.L. and Werry, I. 2005. What is a robot companion--Friend, assistant or butler? In Proceedings of IROS, IEEE, 14881493. DOI= http://dx.doi.org/10.1109/IROS.2005.1545189.
[7] Dautenhahn, K., Woods, S., Kaouri, C., Walters, M., Koay, K.L. and Werry, I. 2005.什么是机器人伴侣--朋友、助手还是管家？IROS论文集》，IEEE，14881493。DOI= http://dx.doi.org/10.1109/IROS.2005.1545189.

[8] Duffy, B. 2003. Anthropomorphism and the social robot. Robotics and Autonomous Systems, Elsevier, 42, 3-4 (Mar. 2003), 177-190. DOI= http://dx.doi.org/10.1016/S0921 .
[8] Duffy, B. 2003.Anthropomorphism and the social robot.机器人与自主系统》，Elsevier, 42, 3-4 (Mar. 2003), 177-190.DOI= http://dx.doi.org/10.1016/S0921 .

[9] Duncan, B., & Murphy, R. R. 2013. Comfortable approach distance with small unmanned aerial vehicles. In , IEEE, (Aug. 2013), 786-792. DOI= http://dx.doi.org/ 10.1109/ROMAN. 2013.6628409 .
[9] Duncan, B., & Murphy, R. R. 2013.小型无人飞行器的舒适接近距离。In ，IEEE，（2013 年 8 月），786-792。DOI= http://dx.doi.org/ 10.1109/ROMAN.2013.6628409 .

[10] Finn, R. L., & Wright, D. 2012. Unmanned aircraft systems: Surveillance, ethics and privacy in civil applications. Computer Law & Security Review, 28(2), 184-194. DOI= http://dx.doi.org/10.1016/j.clsr.2012.01.005.
[10] Finn, R. L., & Wright, D. 2012.Unmanned aircraft systems：Surveillance, ethics and privacy in civil applications.计算机法律与安全评论》，28（2），184-194。DOI= http://dx.doi.org/10.1016/j.clsr.2012.01.005.

[11] Fischer, F. Evaluating enjoyment within AR. Drone Games. A quantitative survey using the Pervasive Game Flow model. 2013. ResearchGate.(Dec. 2015), Retrieved from http://www.researchgate.net/publication/256908477
[11] Fischer, F. Evaluating enjoyment within AR.无人机游戏。使用 "无处不在的游戏流 "模型的定量调查。2013.ResearchGate.(Dec. 2015), Retrieved from http://www.researchgate.net/publication/256908477

[12] Fujii, K., Higuchi, K., & Rekimoto, J. 2013. Endless Flyer: A Continuous Flying Drone with Automatic Battery Replacement. In Ubiquitous Intelligence and Computing, 2013 IEEE 10th International Conference on and 10th International Conference on Autonomic and Trusted Computing (UIC/ATC) 216-223.
[12] Fujii, K., Higuchi, K., & Rekimoto, J. 2013.Endless Flyer：自动更换电池的连续飞行无人机。In Ubiquitous Intelligence and Computing, 2013 IEEE 10th International Conference on and 10th International Conference on Autonomic and Trusted Computing (UIC/ATC) 216-223.

DOI=http://dx.doi.org/10.1109/UIC-ATC.2013.103.

[13] Gleitman, H. 2000. Psychology. Oxford Press. 467.
[13] Gleitman, H. 2000.Psychology.Oxford Press.467.

[14] Goetz, J., & Kiesler, S. 2002. Cooperation with a robotic assistant. In CHI'02 Extended Abstracts on Human Factors in Computing Systems. ACM, New York, NY, 578-579. DOI=http://dx.doi.org/10.1145/506443.506492.
[14] Goetz, J., & Kiesler, S. 2002.与机器人助手的合作。In CHI'02 Extended Abstracts on Human Factors in Computing Systems.ACM, New York, NY, 578-579.DOI=http://dx.doi.org/10.1145/506443.506492.

[15] Gongora, A., & Gonzalez-Jimenez, J. 2015. Enhancement of a commercial multicopter for research in autonomous navigation. In Control and Automation (MED), 2015 23th
[15] Gongora, A., & Gonzalez-Jimenez, J. 2015.增强商用多旋翼飞行器的自主导航研究。控制与自动化》（MED），2015 年第 23 期