I feel being there, they feel being together: Exploring How Telepresence Robots Facilitate Long-Distance Family Communication

Many families often live geographically apart from each other due to work, education, or marriage. Therefore, long-distance families frequently use computer-mediated communication (CMC) tools to stay connected. While CMC tools have significantly improved family communication, they cannot fully mediate social presence. To examine the potential of telepresence robots for improving long-distance family communication, we conducted a two-week qualitative in situ study involving eight families. We analyzed recorded videos of their family interactions and conducted pre- and post-deployment interviews. Our findings highlight telepresence robots’ potential as family communication tools, enabling immersive, natural, and dynamic interactions through physical embodiment and autonomy. Particularly, we identified five categories of family interaction mediated by telepresence robots: engaging in multi-party family communication, exploring home, restoring family routines, providing support, and having joint physical activities. Based on our findings, we present design guidelines for leveraging telepresence robots as effective family communication tools.


INTRODUCTION
Many families live away from each other due to various reasons such as education, employment, marriage, or independence.This physical separation often makes it difficult for families to maintain a sense of togetherness and connection [60].In order to overcome these challenges, many families use computer-mediated communication (CMC) tools to stay connected and keep in touch [26,31,35,62].CMC tools support family communication in various ways, from texting and telephone calls to video conferences and social media.Among them, video calls have emerged as the preferred method of family communication, as they enable richer communication by incorporating non-verbal cues such as body language and facial expressions, facilitating more nuanced interactions [1].Previous research reported that video calls foster closeness and enable family members to have shared experiences across distances [3,45].These tools, however, may only partially mediate social presence, the sense of being together [18,49,54,63].Addressing such constraints of mediating social presence in current CMC tools [49], telepresence robots may offer new possibilities for enhancing long-distance family communication.Through their physical presence and the ability to control movement and proximity [4,32], telepresence robot-mediated communication holds the great potential to foster strong connections and evoke a sense of presence among distant family members.
Researchers have extensively studied telepresence robot use in various public settings, such as healthcare, workplaces, and schools [21,41,44,51,53].However, telepresence robots' potential as a means of mediating family communication has been relatively understudied.While several studies explored the telepresence robot uses for one-on-one intimate relationships [33,46,[74][75][76], there is still a research gap regarding the potential of telepresence robots for remote family communication that involves multiple individuals spanning several generations, including young children and elderly grandparents.In addition, despite home being the primary context of understanding family interactions [65], the role of telepresence robots in mediating family interactions within household spaces has not been extensively researched.As telepresence robots are becoming more affordable even for domestic use (e.g., Ohmni, Ava, Meeting Owl Pro 360, and Kubi Classic) [2,30], they have the greater potential to greatly enhance family communication experiences.Because they are not yet designed for family communication contexts in mind, it will be necessary to explore the potential and challenges of using them for family communication.
Therefore, we aim to address the following research questions (RQs) in this paper: • RQ 1: How do family members use telepresence robots for long-distance family communication?• RQ 2: What are the challenges and opportunities of using telepresence robots for family communication in a home setting?
To address these research questions, we conducted an in-situ qualitative study to investigate how families utilized telepresence robots for daily family communication.Double 2 telepresence robots were deployed in the homes of families whose adult children lived geographically apart from home (See Figure 2).Overall, eight families, a total of 17 local and 12 remote family members, participated in our study over a two-week period.Here, we called a group of participants as 'Local Family, ' those hosting the robot in their home spaces, and 'Remote Family (or controllers),' referring to those remotely controlling the robot.We instructed all participants to record videos from the controller's perspective using a screen recorder each time they used the robot for family communication.We further invited both local and remote family members to participate in pre-and post-interviews to gather their experiences and viewpoints.Postinterviews concluded with the collection of idea sketches to capture design suggestions based on participants' experiences and the challenges posed by existing design and affordance of the telepresence robot.
In summary, this paper contributes to: • Investigating the use of telepresence robots for long-distance family communication within a home space.• Exploring the opportunities and challenges of telepresence robots for family communication.• Presenting design guidelines for improving telepresence robots as family communication tools.

RELATED WORK
We have focused our related work on two major categories: (1) previous research on using CMC for family communication and (2) existing research on telepresence robot uses in various contexts.

Computer-Mediated Communication in Family Interaction
Many individuals live apart from their families for various reasons and often face challenges such as adjusting to new environments, feeling homesick, and missing loved ones [36,60].To bridge the geographical gap and stay connected with distant family members, CMC tools have become essential for supporting long-distance family communication.Families use a range of CMC tools, including video calls, audio calls, text messages, and social media platforms, to communicate with their family members who are physically distant [26,31,35].Previous studies have consistently found that CMC tools are effective in maintaining relationships and fostering connections among distant family members [1,56].Video calls, in particular, have become popular as they enable rich communication by incorporating non-verbal cues such as body language and facial expressions, facilitating richer and more nuanced communication [24].Video calls are particularly beneficial for communication with children who rely heavily on body language to express themselves [3].Consequently, grandparents and other family members often prefer to engage with children through video calls [70].Furthermore, the uses of video streaming for in family communication have increased, allowing remote sharing of significant life events, such as graduations, weddings, and family reunions, in real-time with family members who are physically distant [45].Similarly, 'Always-on' video, a continuous video conferencing of remote family members, enabled distant families to enhance their sense of closeness [50].
Despite this, most current CMC tools cannot fully support the sense of social presence and togetherness [45,62,74].Social presence refers to the sense of being with another person or being in the presence of others, particularly in the context of media use, such as online communication, virtual environments, or remote interactions [6,43].It creates the illusion of physical presence for both the communicator and the recipient by constructing a mental model of a mediated virtual environment [7].Most importantly, this limited social presence often prevents remote families from establishing the same level of connectedness and family interaction through touch, shared contexts, and joint actions [29,60,74], compared to when they live together.Therefore, to understand telepresence, it is necessary to explore how individuals experience a sense of togetherness and how various media interfaces affect this experience.
Despite this, the majority of current CMC tools cannot fully support the sense of social presence [45,62,74].Social presence refers to the sense of being with another person or being in the presence of others, particularly in the context of media use, such as online communication, virtual environments, or remote interactions [6,43].This notion is pivotal in creating an illusion of physical presence for both communicators and recipients by forming a mental representation of a mediated virtual setting [7].However, CMC tools, commonly used in remote communication, struggle to provide sufficient support for fostering social presence, primarily due to a lack of adequate physical feedback and immersive experiences.The lack of social presence frequently prevents remote families from achieving a similar level of connectedness and family interaction, which includes touch, shared environments, and collective actions [29,60,74], as compared to cohabiting families.Some wearable devices aimed to simulate a sense of physical presence for remote family members.For example, Liu and colleagues developed a wearable device that can be worn on the wrist to share and respond to each other's biosignals using animated otter avatars [42].Also, Singha and colleagues proposed Flex-N-Feel, which enables the transmission of touch sensations over a distance using vibrotactile gloves [64].While these devices helped remote families to feel some sense of presence, they primarily focused on allowing remote family members to exchange limited non-verbal cues.
To better mediate social presence and family interactions, we saw the potential of telepresence robots for family communication with their physicality and mobile capabilities.In comparison to conventional CMC tools, telepresence robots have the ability to improve family interactions by providing physical involvement and transmitting a mix of verbal and non-verbal cues, thus enriching family communication [11].Interactions mediated by these robots enable local families to physically engage with remote family members as if they were in the same space, fostering a sense of togetherness.Consequently, our research aims to explore how telepresence robots can contribute to fostering a sense of social presence, promoting togetherness, and supporting social communication among geographically separated families in home settings, leveraging their physicality and mobility.

Telepresence Robot Uses in Various Contexts
Telepresence, defined as the subjective feeling of "being in a remote environment" through a communication medium [67], may greatly enhance long-distance family communication.Researchers identified key factors contributing to telepresence, including behavioral engagement through coordinated actions and sharing of nonverbal behaviors [7], immediacy [71], proximity [9], and agency [34].
Telepresence robots are emerging as an effective tool to enhance the sense of social presence in telecommunications.Telepresence robots offer advantages such as multi-channel communication with the physicality of the robot body and mobility operated by the controller in remote situations [74].Specifically, telepresence robots usually come equipped with cameras, microphones, and proximity controls, allowing them to facilitate fuller multi-channel interaction that could blur the boundary between local and remote participants and robots and remote operators [4,19,39].Previous studies on telepresence robots have focused on public settings like workplaces, schools, and health service facilities [41,44,51,53,68].In these contexts, telepresence robots have demonstrated their ability to enhance social interactions and support a stronger sense of connection among diverse stakeholders.For example, in workplace environments, telepresence robots have enabled remote workers to collaborate and socialize as if they were physically present [41,68].In the field of education, telepresence robots have allowed students to participate in class activities remotely and to interact with colleagues and teachers in a more immersive way [15].Moreover, within healthcare settings, health professionals used telepresence robots to support patients' daily activities and facilitate social interactions, helping them overcome feelings of isolation [38].
Some studies have explored the application of telepresence robots in domestic healthcare and long-distance couple communication.In the context of caring for older adults with dementia, telepresence robots have been used to enhance engagement between family members and the person with dementia [47,48].Additionally, other studies suggest gaze-controlled and hand-controlled accessibility options for people with motor disabilities to facilitate interactions with their friends via telepresence robots, especially in care homes [77,79].These robots have also supported long-distance couples by allowing them to share daily routines [74].Overall, telepresence robots offer interactive and immersive experiences that simulate in-person communication, making them valuable for maintaining connections in long-distance communication.These cases demonstrate that telepresence robots effectively support remote communication by facilitating social interaction and thus enhancing the social presence of remote participants.
In light of the benefits of telepresence robots, it is crucial to explore their potential to improve family communication within domestic settings, with a specific focus on home spaces.Additionally, understanding how these potentials can enhance user experiences and identify design opportunities is essential for optimizing the seamless integration of telepresence robots into domestic environments.However, telepresence robots were rarely investigated in the context of family communication within home settings.As one of the early studies to explore the potential of telepresence robots for family communication, we conducted an in-situ study that included the observations of family interactions via video recordings and interviews involving both local and remote family members.Through this study, we aim to shed light on the opportunities and challenges of integrating telepresence robots for family communication in home spaces.

METHOD
We conducted an in-situ qualitative study by deploying telepresence robots in the homes of eight local families to gain a deeper understanding of their family communication behaviors among geographically separated family members.We recognize the value of in-situ user studies in providing insights into how individuals integrate technology into their daily routines and real-life environments [12,16,58,72].Therefore, we collected recorded videos of their family interactions and conducted pre-and post-interviews with both local and remote family members.All participating families were required to use the telepresence robots for two weeks for family communication and participate in the pre-and post-deployment interviews.Participating families received compensation of 150,000 Korean Won (equivalent to $120 USD) for their participation.

Participants
We recruited eight families, each of which included at least one family member living geographically apart from their local home.In our study, we used snowball sampling, drawing from our social networks and participant referrals, as we encountered challenges in finding family groups willing to adhere to all study procedures through an open call.Some were reluctant to participate in this study, as our study required them to record their family communication and their home spaces, although they could opt out of the recordings they did not want to share.Also, some found it challenging to gain consent for this study from all of their family members.Some groups of people hesitated to use telepresence robots for family communication due to limited digital literacy, especially when only elderly parents were at local homes.Additionally, some families opted not to participate in the research simply because they did not perceive a compelling need for such technology.Because of these aspects of this study, we had to use snowball sampling despite the potential bias it might introduce, such as encouraging more positive experiences or active participation from participants [55].After all, the final participating family groups consisted of individuals who willingly embraced telepresence robot use for family communication.
Within each family, there were two sub groups of participants: the 'Local Family, ' consisting of family members living in the parent's home with access to the telepresence robot (e.g., parents and siblings), and the 'Remote Family, ' comprising family members (e.g., adult children (with their own families)) who had relocated away from the parent's home and remotely controlled the telepresence robot from a distant location.Our study involved eight families, including 17 local and 12 remote family members (See Figure 3 and Appendix A.1).The remote family members left their parent's houses and moved to other cities domestically or internationally.The duration of being apart varied from two months to twenty-two years.The distances between the families ranged from approximately 400 km to 12,000 km.These families represented common scenarios of long-distance family communication for reasons such as marriage, education, and work.Before using the telepresence robot, most participants relied on video and audio calls and family group chats through platforms.Additionally, all of our participating families maintained frequent contact with each other, ranging from at least once a week to nearly every day.

Apparatus
For this study, we chose to use the Double 2 telepresence robot because of its (relatively) affordability and standard features, which included a wide-angle camera, mobile control, and a built-in mic and loudspeaker-features commonly found in other telepresence robots.To be specific, this model was equipped with a 150° wideangle lens, providing a broader perspective of the remote environment.Also, both local and remote families could adjust the robot's height and sound level of the speaker, but the robot does have a restricted minimum height adjustment range, so it could not be lowered below 1.2 meters.Newer and more advanced telepresence robots, such as the Double 3 and AVA, have been released.These models have surpassed the capabilities of the Double 2. They come equipped with enhanced features, such as autonomous driving, click-to-drive functionality, and adaptable scheduling capabilities, which are particularly tailored for work settings.However, these sophisticated features also render them less affordable for home use.In light of this, our choice gravitated towards the Double 2, considering its basic functionality, cost-effectiveness, and alignment with the standard characteristics of typical telepresence robots.This choice was influenced by a desire to reduce potential disparities that could stem from the additional features of more advanced models.By doing so, we aimed to explore how the fundamental features of telepresence robots could be effectively employed for family communication within a domestic setting.
To control the robot and establish connections between local and remote families, specific initiation steps were required.First, remote family members (controllers) were required to download the Double application for remote robot control.Each participating family accessed the application using the provided research account ID and password, enabling researchers to monitor usage frequency and connection duration (See Table 1).Furthermore, local families also needed to turn on the robot and log into the same research account given to remote families.Once both the robot and app log-ins were completed, families could then utilize the telepresence robot for communication.

Procedures
Our study spanned a duration of two weeks, involving three phases: (1) setup and pre-interviews, (2) a two-week telepresence robot deployment, and (3) post-interviews with idea sketches.Interviews were conducted in person in local families' houses while we installed and retrieved the robot.Remote family members participated in these interviews together with local family members via Zoom or FaceTime.All procedures were approved by ethics institutional review boards from the university where this study was hosted, and all participants provided consent to participate in this study.

3.3.1
Set-up and Pre-Interview.Before commencing the study, we scheduled the time for set-up and pre-deployment interviews that both local and remote members could participate in.Then, we brought the Double telepresence robot into the houses of local family members at the scheduled time.Upon obtaining their consent for study participation, we provided instructions to both local and remote members regarding the operation of the telepresence robots, including the fundamental functionalities of the robot.
After completing the setup process, we conducted a joint preinterview session, where both local and remote family members participated in interviews simultaneously.Pre-deployment interview questions aimed to gain insights into their general CMC uses for family communication.For instance, we posed questions such as "How frequently do your family members maintain contact?", and "Which CMC tools do your family predominantly use for communication?"We also asked both local and remote family members about the primary challenges they faced when communicating with each other using current CMC tools.Lastly, we encouraged participants to coordinate their availability to use telepresence robots, especially for families in which remote family members lived abroad due to time differences.

2 Weeks of Telepresence Robot Uses for Family Communication.
For two weeks of deployment, we intentionally did not impose specific requirements or rules on our participants' use of the robot.Our aim was to observe how families naturally incorporated the telepresence robot into their communication.To capture these robot-mediated family interactions, we requested all remote family members to utilize the screen recording feature on their smartphones, tablets, or laptops to record their interaction.These recordings provided comprehensive views of both the local family and controller sides of the family interaction, which helped us understand the events in the home spaces.

Post-Interview including Idea Sketches.
Prior to the scheduled post-interview session, we asked participants to share the recorded videos with us.We explained the purposes for collecting these recorded videos, emphasizing our interest in gaining insights into their natural communication using the telepresence robot.While participants recommended recording all usage instances, they could only selectively share the recordings they felt comfortable sharing.We also encouraged them to review these recordings to illustrate situations that might be hard to describe verbally.We reviewed all recorded videos before conducting the post-deployment interviews, allowing us to identify primary use cases and interesting moments that allowed us to develop tailored interview questions for the post-interview sessions.
In the post-interview sessions, we explored participants' experiences with the telepresence robot over the two-week study period in detail by interviewing all family members.We asked them the overall context of using the telepresence robot, such as timing, location, duration, and reasons for their robot use.We also examined how they used the robot's functionalities in the context of family communication in general.Additionally, we also asked participants to share intriguing and distinctive use cases.Furthermore, by analyzing participants' communication behaviors and patterns using the provided video recordings, we formulated detailed follow-up questions for their primary use cases and interesting moments we observed.We focused on understanding their emotions, satisfaction, and motivations regarding specific situations observed in the recorded videos.
Finally, we examined the challenges and concerns that participants faced while using telepresence robots for family communication.We also encouraged them to provide design suggestions.Participants shared idea sketches based on the pain points they experienced and suggestions for the better design of telepresence robots for family communication.Notably, we received diverse sketch proposals from both remote and local families.Remote family members focused on the system user interface (UI) and the functionality of the control system, while local family members contributed more to the physical appearance and functionalities of the robot itself.

Data Collection and Analysis
We collected a total of 37 video clips, with a combined duration of 629 minutes, from participants.Further, a total of 907 minutes of interview data were audio-recorded and transcribed.For all the data we collected, including video recordings and audio recordings from interviews, we replaced the identifiable information with aliases.All family names used in this paper (e.g., Amy's Family, Belle's Family) were pseudonyms.Moreover, participants were informed that all recorded video data they shared would be securely stored in a local drive accessible only to researchers.The data was permanently deleted after completing the analysis, following the regulations of the Institutional Review Board.
For data analysis, two authors individually analyzed the video clips and interview transcripts and open-coded them.We conducted an initial review of all the video clips to extract and list family activities from notable scenes as textual descriptions.These lists served as a foundation for guiding the analysis of interview transcripts.Further, for the interview transcripts, we conducted thematic analysis [10] to extract codes and themes.The initial codes and themes from each researcher were then aggregated and organized into categorized themes through multiple rounds of iterative discussions among the research team.Throughout this process, some narratives and quotes provided by the participants were incorporated into the main themes to provide supporting evidence for addressing our research questions.

FINDINGS
We categorized participants' use of telepresence robots for interaction with distant family members into five communication behavior categories: (1) engaging in multi-party family communication, (2) exploring family home, (3) restoring family routines, (4) providing practical and emotional support, and (5) having intensive joint physical activities.By examining these activities, we present findings regarding the opportunities and challenges associated with using telepresence robots for family communication in domestic settings.

Engaging in Multi-Party Family Communication
Our participants recognized the telepresence robot to be a practical alternative to conventional video conferencing tools in facilitating remote family communication.Initially, they primarily used the robot for stationary one-on-one conversations, parking it in close proximity.Having been accustomed to the robot, they began to utilize its mobility feature more actively and to engage in conversations involving multiple family members.

Stationary Conversations in a Close
Proximity.The telepresence robot was used to engage in a focused conversation with specific family members.Usually, remote family members who controlled the robot approached the local family members they wanted to talk with and parked the robot near them.During this stationary conversation, we observed that local family members enjoyed the comfort of having physical autonomy, especially their hands, as they did not hold the device but could use their hands freely to express themselves with gestures.For example, Hailey's mother (local family) said that "I relaxed on the sofa while talking to my daughter through the robot.I didn't have to hold my phone or worry about angles, which was very comfortable and supported me to focus more on the conversation.".
Further, remote family members said that they could feel more immersed in remote conversations due to the wide-angle camera.The wide-angle view helped them better observe non-verbal cues from local family members and contextual nuances in the conversation.For instance, Faith (controller) said: "I could easily see what is happening there at home through the wide-angle view, which also made me feel like almost being there." 4.1.2Conversations on the Move.As participating families became familiar with using the telepresence robot, they used its mobility feature more actively to explore the remote environment and interact with multiple family members.We found that local and remote family members experienced different types of autonomy in conversing with mobility.Remote family members felt the autonomy of movement, allowing them to navigate to desired locations and engage in conversations with specific family members.Their movement autonomy allowed them to restore their actual in-person interaction styles with other family members.For example, Ellen (controller) said, "As I moved from the living room to the kitchen, I could stop by my brother's room to say hi.This let us have a natural brother-sister conversation." Belle (controller) also shared her experience when she observed blooming plants together with her mother.She said, "My mom pointed out a blooming plant in the living room.We had bought the plant together when we lived together, and now it had finally blossomed.The ability to explore and experience these things with my parents, rather than just receiving photos or engaging in a fixed-angle video call, made it an exceptional and heartwarming experience" (See Figure 4-D).
On the other hand, we observed that local family members could multitask while conversing with remote family members.They did not have to focus on the communication tool in one location, as remote family members controlled the location of the robot independently.Hailey's mother (local) said, "While I was getting ready to go out, my daughter approached me, and we had a conversation in front of the dressing room.The telepresence robot's ability to move independently meant that I didn't need to hold my phone.It was incredibly convenient to chat with her while I worked on other tasks." Further, the mobility of the robot also helped local family members understand the intentions of their remote family members more clearly, such as where they wanted to see or who they wanted to talk with.David's father (local) shared: "My son turned the robot and followed the cat while talking to me.I realized he wanted to interact with me and the cat.I could sense his intention." Furthermore, we also found that the telepresence robot helped kids to be naturally involved in the conversation.For example, Amy's father (local) said: "Our grandchildren, who previously had difficulty maintaining focus during video calls, wanted to continue the calls and enjoyed interacting with us.They love to drive it!"This gamified nature and mobility features made family communication more engaging for kids, enabling them to participate in family communication more actively.
As such, participants viewed telepresence robots as an advanced alternative to video calls but also noted the need for an intuitive interface and a quick setup process.For family members of all ages, from kids to older adults, to engage in family conversation via telepresence robots, they emphasized that "intuitive controls and more simple ways of setting up and initiating calls for all age groups" (Gina's father, local) would be necessary.
While controllers liked the autonomy of movement, they sometimes found it difficult to focus on conversation because they had to continuously manage robot movement using a keypad while conversing or moving between spaces.Amy (controller) said, "We always need to concentrate on controlling the robot unless it is parked.It would be great if we could control the robot more naturally, such as using voice commands or gestures."Several controller participants suggested expanding control options by reducing the need for a keypad to control movement so they could enhance their physical autonomy, similar to local family sides.

Exploring Family Home
As our remote participants gained autonomy of movement, they started to explore their homes.These explorations allowed them to reconnect with their home spaces and nostalgic family memories.

Reconnecting with Family and Home.
With the autonomy of movement, controllers eagerly explored their homes.This exploration helped them deepen their understanding of their living spaces and foster a stronger sense of connection and belonging to their home and family.For example, Gina (controller) said, "I was thrilled to see my home and reunite with my parents through the robot.I moved around the whole house for a while and even enjoyed an outdoor view with my mom" (See Figure 4-A).This home exploration naturally led to reminiscing about the family memories.In particular, many remote family members found it fascinating to explore the various photo frames and decorations on display in their and (E-F) interacting with local family members while cooking and washing dishes.
family homes.Sometimes, local family members looked around together, sharing their family memories.For example, Amy's sister (controller) said: "I looked at all the photos in the living room with my father.These pictures are like a timeline of my life, from infancy to becoming a mother of two daughters.We reminisced about our memories" (See Figure 4-B).

Rediscovering
Changes in the Home.Some remote family members had been away from their homes for a long time and could not visit due to the COVID-19 pandemic.As a result, for the cases when the local family home was renovated or relocated, telepresence robots gave remote family members a chance to tour their family's new home or renovated spaces.For example, Celine's local family living room was being renovated during the study.Through the robot, remote family members could observe the renovation process by navigating the house through the robot.Celine (controller) said: "I could see the progress of the renovation work in our home and witness the changes" (See Figure 4-C).However, some controllers reported challenges when exploring homes with more diverse layouts and objects.Our participants shared their difficulties of encountering obstacles and limitations when navigating narrow spaces.Therefore, enhancing the robot's mobility and navigation capabilities emerged as a key recommendation.For example, Amy (controller) said: "I would like to emphasize the necessity of versatile wheel mechanisms and control options, such as automatic avoidance technology and gesture controls, in order to address diverse situations and potential emergencies."In ideation sketches, some suggested specific improvements, such as adding wheels designed for navigating two-story houses, implementing obstacle avoidance technology to move around furniture and decorations, and improving maneuverability in tight spaces.

Restoring Family Routines
Our participants used telepresence robots to restore their daily family routines.Using local families' physical autonomy and remote families' mobility autonomy, they could facilitate synchronous family interactions and restore family routines as if they were living together.

Joining Everyday Activities.
We observed that families interacted in various spaces of the homes, particularly in communal areas like the living room and kitchen, which served as common gathering spots for both local and remote family members.For example, families spent leisure time together watching TV, listening to music, exercising together, and having tea time in the living room on weekends (See Figure 5-A,B,C).Remote family members parked their robots near the sofa, actively joining in family routines.Interestingly, Hailey (controller) even selected TV channels and songs through the AI speaker by moving closer to it.Hailey's mother (local) expressed her happiness about her daughter joining their everyday family activities, which made her feel like her daughter was with them.She said: "We watched a baseball game in the living room and sang cheering songs together to cheer for our team through the robot." Remote family members also participated in everyday activities like mealtime and housework in the kitchen.This allowed them to participate in conversations, offer help, and stay connected with their family members' daily lives.For example, remote family members who lived alone joined local families' mealtimes.Controllers moved the telepresence robot to the kitchen table and shared a meal.Belle (controller) said: "I parked the robot on the kitchen table and had dinner together.As we doing the same activity, I could make eye contact with my mom, so I felt like I was there" (See Figure 5-D).This arrangement allowed remote family members to participate virtually in the shared dining experience and feel a sense of togetherness despite being physically separated.Moreover, some local members interacted with their remote family members while performing household chores like washing dishes and cooking (See Figure 5-E,F).For example, Faith's mother (local) mentioned, "When I cooked dinner in the kitchen, my son would come to me and ask about the dinner menu and recipes." Our participants also restored shared activities in more diverse spaces, such as the front door, hallway, bedroom, and dressing room.For example, remote family members could greet their local family members when they arrived home by moving to the front door.For example, Amy's sister (local) shared her experience of going to the front door to welcome their mother back from the gym and meeting her sister in the hallway on her way to the front door, feeling like "I really liked being there beside my sister and mother."Also, Amy (controller) spent much time in the dressing room and bedroom with her mother and sister.It provided them with an opportunity to share clothing purchases and help their younger sister dress for upcoming meetings and dates.Amy's sister (local) said: "My sister and I used to spend much time together in the dressing room when we lived together.
[Through the telepresence robot], we can now reunite and engage in this activity again." Moreover, Celine (controller) said she chatted with her mother and sister in the bedroom before bedtime as if they were to live together.For example, Celine (controller) said "I used to talk a lot with my sister and mother.It always happened when we lived together.I am glad we can restore this again." As such, various family rituals and routines took place throughout the entire home.However, there were few privacy concerns reported with regard to the use and access of telepresence robots.Local families willingly allowed remote family members access to any place in the entire home space.Celine's sister (local) said: "We have lived together for more than 20 years, so there is no need to protect privacy for our own family members."

Participating in Family
Gatherings on Special Occasions.In addition, remote families used the telepresence robot to participate in family events and celebrate special occasions.For example, on Lunar New Year's Day, remote family members could greet the whole family by bowing politely through the robot.They could also join the ancestral rites.Celine's mother (local) said, "My daughter's family bowed politely and connected with us during the ancestral rituals.It was a special memory for us that they could join and be emotionally present with us." In another case, family members celebrated a remote family member's birthday with the robot (See Figure 6-A).Amy (controller) said, "I could celebrate my birthday with my parents and sister through the robot.I could see how much they celebrated my birthday, and I felt we were together."Additionally, Ellen's family shared their experience of a big family gathering: "When our in-laws visited our home, my daughter participated in our family gathering via the robot.She could interact with her grandparents, whom she missed a lot" (See Figure 6-B).
Remote family members also noted that the robot gave them more opportunities to participate in various family activities, creating a strong sense of telepresence.Ellen (controller) shared: "I have been living alone in Hong Kong for over three years due to my work, and I have missed my family tremendously.So, when I saw all my family members on one screen, I felt like I was right there with them, and it helped alleviate my homesickness."Also, local family members said that they sense the presence of remote family members through the robot's physical embodiment.Some participants, however, felt there was room for improvement in fully restoring family routines through telepresence robots.First, some participants shared some challenges in initiating and setting up the phase.Many of them mentioned the inconvenience of the multiple steps required to connect the robot, which could disrupt their natural daily routines.For example, Gina's sister (controller) said: "We want to use this robot daily, but it's too complicated for us during the initiation stage.We always have to ask my local family for assistance in turning on the robot and the app to receive my call." Therefore, our participants suggested simpler access methods, such as password input and automated scheduling, and proposed modes like 'always on.' Regarding the always-on mode, Gina's father (local) added: "We have no privacy concerns with our daughters; they should be able to connect freely without our permission." In addition, while most families had minimal privacy concerns among family members, they were more concerned about security from external threats.Some participants were worried about potential camera hacking, so they suggested a removable camera cover to protect family privacy when the robot was not in use.Celine's sister (local) mentioned: "We are concerned about hacking or camera breaches.A removable camera cover, like laptop webcam stickers, would be very helpful to protect our family privacy from others." Moreover, they said multi-remote family engagement presented challenges.As the current telepresence robot system allowed only one controller at a time to join the family interaction, there were some difficulties if there was more than one distant member, like David, Faith, and Gina's cases.Therefore, these families suggested implementing a multi-remote user system that enabled several controllers to engage in family routines.
Further, some also proposed displaying contextual information on the system screen, such as local time and weather.These suggestions aimed to enhance mutual understanding among family members.For example, Ellen (controller) said: "Displaying time differences would be helpful for enhancing mutual understanding between local and remote family members in everyday life" (See Figure 7-A).
Finally, some participants suggested the potential for additional features that could add extra value to a home environment.Therefore, some participants suggested integrating smart home devices and systems into telepresence robots (See Figure 7-B).They suggested features such as remote control of lights, thermostats, and other smart devices through the robot's interface.Furthermore, some participants also proposed adding housework functions to the telepresence robot, such as vacuuming and wet mopping capabilities (See Figure 7-C).Celine's mother (local) said: "With the ability to help with cleaning tasks and explore the house, a robot could be more valuable and create win-win situations, such as assisting parents with cleaning while also allowing remote family members to engage with the home environment."

Providing Practical and Emotional Support
As telepresence robots facilitated synchronous and immediate interaction, some families used telepresence robots as support tools to provide practical and emotional assistance.

Supporting Practical Assistances in Sync.
We observed that some participants assisted elderly parents or pregnant daughters who needed immediate support.For example, Gina's father (local) asked her daughter (controller) for help with tasks like solving technical issues or reading small fonts in manuals due to vision problems.Gina's father (local family) said: "I asked my daughter to read the small font in a manual.She came over, positioned the robot, and read it aloud for me" (See Figure 8-A).In addition, Belle (controller) said, "I got wrist pain after being pregnant, so my mother sent me the wrist guard.She guided me step by step on how to wear it.It was immensely helpful for obtaining information and seeking direct assistance" (See Figure 8-B).

Fostering Emotional
Connections with Non-Verbal Cues.We also observed our participants establishing strong emotional connections through nonverbal interactions like haptic feedback and facial expressions in family interactions.Some participants engaged in haptic interactions, including hugging the robot or tapping its head to express their affection.In particular, Belle (controller) mentioned how the telepresence robot helped nurture a strong bond between the baby in the womb: "I could share the changes in my body with my mother, who monitored my well-being and offered practical guidance.Moreover, I would show my baby bump on the robot's screen, allowing my parents to interact with the baby in the womb.My mother even read the book for the baby in the womb.Seeing their loving facial expressions showed just how much they cared for the baby." However, these participants also shared some suggestions regarding the current appearance of telepresence robots as a support tool.When it came to emotional support and information, they were dissatisfied with the rigid and mechanical robotic form that lacked human qualities.Therefore, they suggested adding more human-like and customized features that would enable the robot to better emulate human body features (See Figure 9-A).For example, Ellen's mother (local) suggested: "Adding a hanger to the robot's appearance will allow us to dress it up more like my daughter." Moreover, in terms of physical capabilities, many participants suggested adding hands or arms to robot bodies, which would allow for more nuanced interaction and practical assistance (See Figure 9-B).For example, Faith's mother (local) said: "The robot should have more human-like functions with delicately controllable arms and hands, allowing for more precise interactions."They also wanted adjustable height options to accommodate interactions with babies or pets and the ability to control the robot's head angle to emulate bowing and nodding (See Figure 9-C).David's father (local) said: "Having more adjustable height and head angle control would be helpful for maintaining eye contact in various postures, including when interacting with babies or pets." They perceived that implementing these enhancements could significantly improve the robot's ability to provide a more natural and comprehensive form of support.

Having Intensive Joint Physical Activities
As our participants adapted to telepresence robot-mediated communication, they gained an in-depth understanding of the robot's functionalities and improved their control skills.Therefore, this led some families to use the robot for more intensive physical activities.In particular, we observed that interactions with kids and pets exhibited more physical interaction and dynamic movement than those with other adult family members.
For example, Amy's twin kids (controller) played hide-and-seek with their aunt (Amy's sister; local) using the robot (See Figure 10).Amy's sister (local) said, "My twin nieces are eight years old, and they enjoy playing with me.We played hide-and-seek using the robot.I hid somewhere in the house, and my nieces controlled the robot to explore the home and find me.It was a fascinating experience."In addition, some remote members interacted with the local family's  pets by moving around to find them.The mobility of the robot made them feel like they were playing with their pets in the same place.For example, David (controller) said, "I searched around the whole house to find my cat, and I found him under my old bed.It feels like we are connected and can sense each other." However, these participants faced challenges with the robot's limited battery capacity, as this issue became more prominent during physically active activities that required more frequent movements.They shared instances where the robot ran out of battery while engaging in joint physical activities.While the manufacturer said that Double 2 generally had 3 hours of battery life, frequent control and extensive movement drained the battery faster and caused unexpected halts.Therefore, David (controller) suggested that: "The robot should extend the battery life and alternative charging methods such as wireless charging to streamline the charging process."

DISCUSSION
In this paper, we investigated how telepresence robots facilitated long-distance family communication within a domestic environment.While previous research has primarily focused on the use of telepresence robots in public settings and one-on-one interactions between intimate relationships [41,44,51,53,74], our study sheds light on their potential of telepresence robots as an advanced family CMC tool.In particular, we identified five unique categories of family communication and activities that telepresence robots could offer for communication between remote and local members across generations: "engaging in multi-party family communication", "exploring family homes", "restoring family routines", "providing practical and emotional support", and "having intensive joint physical activities".Initially, most of our participants just perceived the telepresence robot as an alternative to video calls, so they predominantly engaged in one-on-one stationary conversations, parking the robot in close proximity to one of their local families.However, as they became more accustomed to using the robot, they expanded their uses of the robot into mobile multi-party family interaction.Furthermore, mobility control types varied among family members.In a previous study, the researchers found that grandparents tended to stay in a single fixed location and experienced a burden on multitasking, such as moving and engaging in activities simultaneously-a cognitively demanding task [52].However, in our study, when children/grandchildren controlled the robot, they tended to engage in more intensive physical movement and diverse activities, moving around the local home.Overall, we observed that telepresence robots have the potential to facilitate more immersive, natural, and dynamic family interactions.
Our findings also underscored the differing sense of agency and presence experienced by remote and local family members.For local family members, direct interaction with the robot provided a tangible representation of their remote family members.This physical embodiment, as well as the physical autonomy that local families were not required to hold the device or to remain stationary, enabled natural everyday interactions that replicated in-person family interactions.Therefore, these elements fostered a profound sense of social presence, feelings of "being together" [13,37] with their remote families.On the remote family side, the autonomy of movement led them to have more immersive and dynamic family interactions in their local home by allowing them to move to where they wanted to go and engage in family activities.These features fostered their sense of telepresence, the sense of "being there" in their local home environment [27].
An intriguing observation from our study was the minimal privacy concerns reported.This finding was likely influenced by the placement of the telepresence robot in the parents' home while being controlled by the children.Such a setup might lead to fewer issues regarding privacy.In contrast, a previous study revealed significant privacy challenges when the robot was placed in the homes of children or grandchildren.Younger family members often preferred to keep certain areas of their homes private from their parents or grandparents [52].Our findings implied that the parent's local home was not just viewed as the parents' living space but also as a communal area with shared memories and history.As a result, we suggest that privacy concerns in the context of telepresence robot use could vary based on generational viewpoints and the specific location of the robot's deployment.
In addition, there were some challenges posed by the fact that most telepresence robots are not designed primarily for purposes of family communication in mind.In the following section, we present design implications aiming to enhance the utility of telepresence robots for facilitating family communication within domestic settings based on the challenges and suggestions from participating families.

Design Implications
In this section, we propose six design implications, highlighting ways to enhance the utility of telepresence robots as family communication tools (See Table 2).

Facilitating Flexible Control of Movement.
Our study identified challenges in controlling collisions with furniture and navigating tight spaces and different floors within homes.To address these issues, we propose improving the robot's navigation system.This includes developing algorithms for path planning and obstacle recognition to support social capabilities [17,61,80], alongside equipping the robot with sensors like depth cameras and proximity sensors to avoid collisions in real-time [20,25,28].
Furthermore, our study revealed a desire among remote family members for more intuitive control options to better concentrate on family interactions.To facilitate easier control of the robot, we recommend the integration of voice and gesture-based control mechanisms.Such controls would offer a more natural and seamless interaction experience, thereby enriching the engagement of both local and remote family members in their interactions, as suggested by prior work [77,79].Also, this could ensure inclusive family interactions that accommodate family members with special needs.
Finally, our participants expressed concerns about battery life, particularly in the context of intensive physical activities.Therefore, we propose improving battery management through the integration of energy-efficient components and increased battery capacities.
5.1.2Simplifying the Initiation Process for Daily Use.In finding 4.3, some participants felt that the initiation process to connect with each other via the telepresence robot was overly complicated.This complexity may disrupt their ability to use the robot seamlessly within their natural daily routines.Furthermore, we observed that privacy concerns between local and remote family members were generally low, whereas concerns about external threats were high, which was consistent with findings in other contexts [40,59,78].To address this, we propose simplifying access settings between local and remote family members.For example, simplified access modes, such as 'always-on' and 'do not disturb' modes to protect minimal privacy, could be helpful.Additionally, to address potential external threats to privacy, we recommend equipping telepresence robots with detachable or closable cameras to safeguard user privacy in domestic use.

Supporting Multi-Remote User Scenarios.
As highlighted in finding 4.3, some families even have more than one remote family member.To accommodate these dynamics, we recommend a multi-remote user interface that could allow concurrent family interactions involving multiple remote participants.However, some challenges may arise from managing UI complexity and conflicts during simultaneous control.To address this, we suggest implementing visual or audio cues to coordinate and identify the controller in multi-user scenarios.

Implication Guideline
Facilitating flexible control of movement • Enhance navigation for various home layouts and obstacles.
• Incorporate voice and gesture control for the controller's autonomy.
• Support accessible controls to accommodate family members with special needs.
• Explore wireless charging and optimize power management to extend battery life.
Simplifying the initiation processes for daily use • Offer flexible access options with time setting like 'always-on' or 'do not disturb' mode.
• Use detachable camera covers to protect family privacy from external threats.
Supporting multi-remote user scenarios.
• Recommend a multi-remote user interface that could allow concurrent family interactions involving multiple remote participants.
Considering the versatility of telepresence robots • Integrate telepresence robots with smart home tools enabling remote control of appliances and the execution of physical housework tasks.
Improving user interface for multigenerational family interactions • Enhance the user interface to make it easier to use and more intuitive for users of all generations.
Strengthening human-like appearance and functions • Allow customization of the robot's appearance.
• Integrate hands and arms and more adjustable height settings for dynamic interactions.
Table 2: A summary of the design implications for improving telepresence robot using for family communication in domestic use.
systems to enhance immersion in remote locations and transform the robot into a practical home tool.Building on a prior study that aimed to integrate distant partners' lives with voice-activated smart home devices [75], we suggest integrating the telepresence robot's system with various smart home tools, enabling remote control of appliances and the execution of physical housework tasks.We believe that this integrated system can merge the daily lives on both families' sides, allowing remote families to actively engage in the daily routines of their local families and cultivate a profound sense of belonging.This versatility of telepresence robots may facilitate their widespread adoption as a family tool.
5.1.5Improving User Interface for Multi-Generational Family Interactions.In a family group with family members spanning multiple generations, adapting to controlling the robot could be challenging.As a result, we suggest enhancements to the user interface to make it easier to use and more intuitive for users of all generations.For instance, these enhancements may include visual cues, such as labeled buttons [14], infographics to indicate time differences and the incorporation of a speed control gauge.Further, to address safety concerns, a warning signal could be implemented to alert when recommended indoor speeds are exceeded, particularly for children driving [5].This will be particularly beneficial for young children and older adult users who may have difficulty initiating and controlling the robot.
5.1.6Strengthening Human-like Appearance and Functions.Based on finding 4.4, some participants expressed their dissatisfaction with the current robots' simplistic appearance.As previous studies advocating friendly robot appearances and personalization through clothing [22,69], we propose allowing users to customize the robot's appearance with features like hangers and dress-up options for a closer resemblance to distant family members.This approach has the potential to nurture stronger emotional connections and enhance the sense of presence during family communication facilitated by the robot.Furthermore, building on previous studies that have investigated the incorporation of arms into telepresence robots for collaborative gestures and physical actions [8,66,73], we believe that equipping the robot with human-like arms and hands will be a crucial improvement for enhancing its capabilities for family CMC tool.This addition may allow family members to engage in more nuanced and practical interactions during the telepresence robot-mediated family interactions.In addition, our finding 4.5 revealed limitations in engaging with children and pets, especially the challenges related to the restricted adjustability of the robot's height function.To enhance interactions for all family members, we recommend expanding the range of height adjustment options, given the emphasis of prior studies on the impact of height on awareness and communication processes [19,23,57].

LIMITATIONS AND FUTURE WORK
Our study offers valuable insights into the role of telepresence robots in facilitating long-distance family communication, shedding light on both opportunities and challenges.However, it is important to acknowledge several limitations, which we plan to address in future research.
First, we employed a snowball sampling method to recruit participating families, which may introduce potential bias.Additionally, our participants generally possessed higher levels of technical literacy and open-mindedness toward new technology, which may limit the generalizability of our findings.
Second, our study was conducted within the Korean context, and its applicability to other cultures or backgrounds may be limited.To enhance the generalizability of our results, future research should employ more diverse sampling methods, encompassing individuals from various cultural backgrounds and increasing the number of participating families.
Third, the duration of our study was limited to two weeks, and it was conducted within the environment of the local family.For future research, it would be beneficial to consider extending the duration of the study.This extended timeframe would allow for a more comprehensive investigation into how family routines and behaviors might develop and change over a longer period of telepresence robot deployment.Such an approach could provide deeper insights into the long-term impacts and adaptability of families to this technology in their daily lives.
Fourth, it would be insightful to reverse the study setting, where remote family members (such as children) act as local hosts while the local family members (like parents or grandparents) operate the telepresence robot.This shift could offer valuable perspectives on privacy concerns in family communication from various angles.
Fifth, with the advent of newer telepresence robots (e.g., Double 3, AVA) that offer advanced features, further investigation would be necessary.Future research would need to focus on assessing how these improved functionalities influence family communication and usage in home environments, thus enriching our understanding of the evolving role of telepresence technology in domestic contexts.

CONCLUSION
In this study, we delved into how telepresence robots can enhance long-distance family communication within a home environment.Our approach involved conducting an in situ qualitative study, encompassing the collection of recorded videos capturing family interactions and pre-and post-interview data from eight participating families.As a result, we identified five categories of family interaction mediated by telepresence robots: engaging in multiparty family communication, exploring the family home, restoring family routines, providing practical and emotional support, and having intensive joint physical activities.Our findings underscore the potential of telepresence robots as advanced tools for family communication, facilitating more immersive, natural, and dynamic interactions.Interestingly, we observed differing senses of agency and presence between local and remote family members.Local families reported a strong sense of "being together" attributed to physical embodiment and physical autonomy, while remote families expressed a sense of "being there" characterized by immersive and dynamic interactions with movement autonomy.These diverse experiences have led us to identify opportunities, challenges, and potential enhancements for telepresence robots in the context of family communication within home settings.We hope our findings can offer valuable guidance for researchers and designers.

Figure 2 :
Figure 2: Example of interacting via telepresence robots in family homes: local family members engage with their remote family members through the use of a telepresence robot.

Figure 4 :
Figure 4: Examples of exploring home together with local families or on their own: (A) viewing the outside scenery together; (B) looking at photo frames; and (C) observing the renovated living room.

Figure 5 :
Figure 5: Examples of restoring family routines in the living room and kitchen: (A) joining tea time; (B) watching baseball games; (C) doing exercising together; (D) having lunch together;and (E-F) interacting with local family members while cooking and washing dishes.

Figure 6 :
Figure 6: Examples of restoring family routines by participating in family events: (A) celebrating birthdays together; (B) attending big family gathering.

Figure 7 :
Figure 7: Examples of ideation sketches for improving telepresence robots at a user interface (UI) level: From (A) to (C) suggested by Ellen, Faith, and Celine.

Figure 8 :
Figure 8: Examples of providing practical support for distant families: (A) reading fine print in manuals for elderly parents; (B) explaining how to use a wrist guard to a pregnant daughter.

Figure 9 :
Figure 9: Examples of ideation sketches for improving telepresence robots at a physical level: From (A) to (C) suggested by Ellen, Faith, and David's family.

Figure 10 :
Figure 10: Examples of playing hide-and-seek games using a telepresence robot: (A1-3) demonstrate how a local family locates the robot; (B) illustrates how a remote family finds their family member at home.

5. 1 . 4
Considering the Versatility of Telepresence Robots.According to finding 4.3, our participants desired more household-friendly functions in telepresence robots besides communication tools.They recommended integrating with smart home devices and housework

Table 1 :
Usage data include average call duration, total two-week use, and the number of times used (names anonymized).