Humanoid robots for assisting people with physical disabilities in activities of daily living: a scoping review

ABSTRACT

The aim of this scoping review was to gather, summarize, and map the knowledge of peoples’ experiences on humanoid robots, capable of assisting people with activities of daily living. The review was guided by the framework of Joanna Briggs Institute and PRISMA-ScR. We found 44 papers describing nine robots that could assist with a variety of tasks listed in the International Classification of Functioning, Disability, and Health. The mixed population experienced little or no anxiety toward the robots, and most accepted the robots’ ability to monitor for safety reasons. Some participants disliked the robots’ large size and slow movements. Most of the participants found the robots easy to use. They wanted improvements in the robots’ functionality and the ability to personalize services. Several of the participants found the services interesting and useful, but not for themselves. The experiences of humanoid robotic assistance showed an insufficient level of technical readiness for assisting in physical assistance, a lack of personalization and readiness for use in home settings. The practical relevance of these findings lies in guiding future research and development toward a more individualized approach focusing on user needs and experiences to enhance the efficacy and integration of humanoid robots in health-care.

KEYWORDS:

• assistive robot
• disabled
• independent living

Introduction

Having a physical disability greatly affects the independence for many people, as they may be in more or less constant need of assistance from both formal and informal caregivers. In 2030, there will be an estimated shortage of 5.7 million nurses worldwide. This shortage of resources is partly demographically driven by population growth and aging. Additionally, it is linked to the ambition goals of service delivery outlined in the context of the Sustainable Development Goals (Broadbent et al., 2016; World Health Organization, 2017, 2020). The future strain on health services might reduce the help for people with physical disabilities who need assistance with their daily activities (ADL).

Recent technological advances have aimed at the provision of services facilitating self-management and autonomy, which pertains to an individual’s capacity to exercise personal choices and maintain control over their lives. This progress particularly targets individuals with diverse disabilities, aiming to enhance their independence, while alleviating burden on healthcare systems (Gignac & Cott, 1998; Kritikos, 2018; World Health Organization, 2020). Such technologies may include memory aids, sensors for detecting falls, and safe delivery of medication. In addition robots (humanoid and non-humanoid) have emerged as a healthcare technology that offer independence and safety for users, as well as high quality and consistent services (Pripfl et al., 2016b; Reeves et al., 2020; University of Cincinatty UC Health, 2021).

For example, social robots such as Pepper and Nao (Robinson & Nejat, 2022) may provide companionship, reminders for cognitive disabled and supervision of simple exercises. Both Pepper and Nao are humanoid robots developed by Softbank Robotics. Pepper is larger, more advanced and designed for commercial applications, while Nao is smaller, simpler, and commonly used in research and educational settings (Softbank Robotics, 2023). Studies show that such robots are experienced as enjoyable, friendly, and safe (Bilyea et al., 2017; Melkas et al., 2020; Papadopoulos et al., 2020). They are however not capable of assisting users with physical tasks, such as bringing a glass of water or picking up something from the floor. Examples of available assistive robot products for people with disabilities are My Spoon (Song et al., 2013) that can assist with eating, and JACO (Beaudoin et al., 2019) that can assist users in gripping objects. Robots able to assist users with a variety of physical tasks such as Hobbit, Care-O-bot and PR2 (Table 5), have been tested in trials but are not yet fully developed to be used as home care assistants for people with disabilities. Such robots could contribute to independence for this potential user-group. This includes individual`s dependent on human carers, such as those who have suffered a stroke, spinal cord injury, cerebral palsy, or paralysis due to other neurological conditions. However, as for all welfare and health technology products, such robots will need to be accepted by users. Low acceptance of technologies has been associated with non-adoption or abandonment of the services (Heerink et al., 2010; Venkatesh et al., 2003). Investigating peoples’ experiences with such robots is therefore essential.

Previous reviews have investigated different users’ perceptions and opinions of assistive robots. Vandemeulebroucke, et al (2018a, 2018b, 2021) investigated older, mainly healthy adults`, experiences, and opinions of a wide range of aspects of socially assistive robots. The participants had both negative and positive experiences of the robots assisting them in daily life. The authors suggested that investigating lived experiences can create an opportunity for technological innovations in healthcare. Savela et al. (2017) found more positive experiences than negative in studies of attitudes toward robots in elderly care. Papadopoulos et al. (2020) investigated enablers and barriers for implementation of socially assistive robots in health care. The enablers found were enjoyment, usability, personalization, and familiarization. Barriers were related to technical problems, to the robots’ limited capabilities and the negative preconceptions toward the use of robots in healthcare. All these reviews report a summary of results from a variety of studies, including a range of different robots, both assisting in terms of prompting and reminding, providing entertainment but only a few, assisting with physical assistance. However, there is a need to comprehend the current status of humanoid robots able to assist physically disabled people in ADL, and perceived experiences of humanoid robots assisting in ADL. This knowledge is crucial to assess the potential of robots for supporting the autonomy and independence of physically disabled individuals as well as alleviating the burden on healthcare systems.

Objective

This study aimed to give an overview of the robots capable of physically assisting in a variety of ADL activities, like fetching and picking up objects, to enhance the independent living of people with physical disabilities. Currently, none of the robots providing physical assistance appear to be available on the market. However, given the ongoing evolution of technology our goal was to explore whether these robots have undergone testing and to understand people’s experiences with them, either through direct interaction or video demonstrations. The research questions guiding this review were:

1. What kind of humanoid robots exists that can physically assist in ADL (such as bringing and picking up objects) and potentially support the independent living of people with physical disabilities?

2. What are users’ experiences of humanoid robots assisting with ADL in terms of; perspectives of the robots’ efficacy in aiding in ADL, the visual aesthetics of the robots, users’ perceptions of safety and privacy and the ease of operating the robots?

We aimed to gather, summarize, and map the range of knowledge to understand the status of the novel research field on robots for assistance in the home. We aim to fill a gap in the literature on this topic, as there is currently sparse knowledge in this area. The insights gained from this review are anticipated to offer novel contributions to the field, providing valuable guidance for stakeholders including robot developers and healthcare and welfare technology service providers.

Method

Study design

Preliminary searches conducted within various research databases indicated a limited research field pertaining to humanoid robots assisting with physical ADL activities. This exploratory phase yielded a diversity of study designs and studies with early stages of research. Consequently, we adopted a Scoping Review methodology as the aim was to provide a broad overview of the literature. A Scoping Review is appropriate when exploring a complex or emerging field (M. D. J. Peters et al., 2020).

We used the Joanna Briggs Institute (JBI) framework (M. D. J. Peters et al., 2020), and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) (Tricco et al., 2018) to guide the rationale for the review, eligibility criteria, information sources and selection of evidence and how to report the results. PRISMA-ScR was used to ensure transparent and complete reporting (Tricco et al., 2018).

Our study protocol was registered in Current Research Information System in Norway; Cristin project ID: 2526318.

Data sources and search

We searched CINAHL, MEDLINE, EMBASE, APA PsychInfo and Scopus. The different databases were searched for eligible studies from 2012 to 2022. The search strategy, including all identified keywords and index terms, was adapted for each database and/or information source (Table 1). Primary search terms were described by using the Population, Concept and Context elements (M. Peters et al., 2021) (Table 1). Each of the terms defining the context were paired with the terms defining the concept of robots, activities of daily living and experience. We did not specify terms for the population as pilot-searches showed that this restricted the search and results considerably. An experienced librarian supervised the search strategy based on the inclusion and exclusion criteria (Table 2).

Table 1. Primary search terms.

Concepts	Primary search terms
home context	nursing home*, long term care, homebound*, home* living, independ*, commut*, communit*, district*, municipal* residential*, dwelling, adl, self* assistive
robots	robot*, robotics, care-o-bot, tiago, lio, eve, pr2, molly, max, Charlie, era, agent*, care*, companion*, service*, humanoid*, assists, assistance, assistive, home, hobbit, health, healthcare, support*, domestic, collaborat*, cobot*
activities of daily living/care	daily life activity, daily life living, home care, Home Care Services, self-help device, Activities of Daily Living+*, Independent Living, community living, Self-Care Skills, Daily Activities, Assistive Technology, Assistive Technology Devices, Self Care, Self-Care, Self-Care Skills
experience	experience*, attitude*, perception*, satisfact*, view*, belie*, opinion*, perspectiv*, usefulness*, usabil*, accept*, perception*, need, needs

Table 2. Inclusion and exclusion criteria.

	Inclusion	Exclusion
Population	People of all ages	Healthcare personnel, Caretakers
Concept	A Humanoid robot able to assist physically in ADL activities- as defined by ICF: mobility, self-care, and domestic life	Robots for; Surgery, pain-management, rehabilitation, exoskeletons, cognitive assistance/training, social skill development, managing symptoms, brain functions, monitoring, communication/telepresence. Technical robot features (e.g., navigation), assistive devices, robot-pets
Context	Users’ homes, Home, Care facility, nursing home, simulated home facility	Hospital

Abbreviations: ADL = Activities of daily living, ICF = International Classification of Functioning, Disability and Health.

Inclusion and exclusion criteria

We included humanoid robots with human features, such as arms and fingers. These body-parts were seen as necessary to assist persons with physical tasks in a home setting (ADL activities). We did however apply a broad definition (Cambridge Dictionary) of humanoid robots when discussing discrepancies between authors: “a machine or creature with the appearance and qualities of a human.,” to ensure that we did not miss out on any important studies.

In addition we used the Classification of Functioning, Disability, and Health (ICF) and the domains of self-care, mobility and domestic life (World Health Organization, 2021), as a framework for ADL activities. The ICF is the World Health Organization’s framework for measuring health and disability at both the individual and population levels. This meant, we included robots able to assist in i.e. (but not limited to): eating, drinking, washing oneself, transferring and tidying or cleaning.

Due to the rapid development of this technology, we included research papers, conference papers/proceedings, book chapters and reports (Levac et al., 2010). The papers had to be written in English. The preliminary searches revealed that the relevant papers on the targeted robots were within the past decade. Consequently, we incorporated papers from the last 10 years.

Exclusion criteria were guided by our focus and research questions. We therefore excluded robots with no human features, and those used exclusively for monitoring, communication, companionship, and robots for surgical purposes. We excluded assistive devices for ADL. We also excluded studies in which participants were not exposed to any robot either directly or by video.

Data extraction

As suggested by Arksey and O’Malley (2005) and Levac et al. (2010), the team developed a standardized list for screening after discussion of the preliminary findings from the pilot searches. The search was performed in January 2022 and resulted in 8017 studies. First, 3197 duplicates were removed by automation tools or manually (The EndNote Team, 2013). Then, 4820 publications were transferred to Rayyan software for screening of the papers (Ouzzani et al., 2016).

Data highlighting the research questions were extracted from the retrieved papers and transferred to excel under the following elements: author(s), year of publication, country of origin, study aim, population, method, context, robot name, connection to a project and participants experiences of the robot.

Transparency

Two of the co-authors independently screened the papers by title and abstract and met regularly to discuss conflicts. Inter-rater discrepancies were resolved by discussion with the third author. Full-text studies were screened by two independent authors, with a second discussion on the discrepancies with the third author. The reference list was then screened for additional studies. In addition, we searched the web and contacted authors of the included papers and distributors of the robots for information about the status of the robots.

To ensure that the extraction process was rigorous and transparent, the standardized data abstraction tool was pilot tested among all reviewers. The data was extracted by the first author, discussed for consensus by the second and third author.

Results

Figure 1 presents the PRISMA flow diagram (PRISMA-ScR) (Tricco et al., 2018). The result was 44 full text papers. The reasons for exclusions are listed in the flow diagram.

Figure 1. Flow diagram for study selection.

The papers, published between 2012–2022 originated from Austria, Belgium, Germany, Greece, Israel, Italy, Japan, the Netherlands, Poland, Sweden, the UK and the USA. They encompassed both journal and conference papers and were published in journals on health and technology. Many of the papers presented both the development of the technology, and users’ experiences of the robots. Among the papers, seven employed quantitative methods, 10 utilized qualitative methods, and 28 adopted mixed methods. Some questionnaires that were used were inspired by established models such as the technology acceptance model (Venkatesh & Bala, 2008), The unified theory of acceptance and use of technology (UTAUT) (Venkatesh et al., 2003) and the Almere model (Heerink et al., 2010). Standardized questionnaires applied included the Godspeed (Bangor et al., 2008; Bartneck et al., 2009; Piasek & Wieczorowska-Tobis, 2018), the system usability scale (SUS) (Bartneck et al., 2009; Bevilacqua et al., 2015; Cavallo et al., 2018; DiNuovo et al., 2018; Fischinger et al., 2016; Körtner et al., 2014), the user experience questionnaire (Schrepp et al., 2014) and the AttrakDif (Gerlowska et al., 2018; Hassenzahl et al., 2003).

The final sample of papers, shown in Table 3, include author(s), year of publication, aim of study, population, method, context, and users’ experience with robots.

Table 3. Included papers with aims, population, robot, method, and participants experiences on robot interaction, appearance, functionality, and usability.

Author Country of origin	Aim	Population n=, gender, age	Robot	Method Interaction/video Location	Participants` experiences (1) Interaction (2) Apperance (3) Functionality (4) Ease of Use
(1) Arnold and Scheutz (2018) USA	Explore how a positive or negative attitude from the robot and whether the robot touches an interactant, affects how a robot is judged as a worker and teammate.	332: 135 female, 197 male. Av age 44.	PR2	Quantitative Online Survey with Video.	Rated a robot that initiated touch higher on skill, attentiveness to needs and desires, being a desirable teammate, fairness, and values. Female respondents more positive to a lack of robotic touch.
(2) Bajones et al. (2020) Austria (Greece, Sweeden)	Explore if older adults experience the HOBBIT robot as a suitable means to maintain independent living in their private household.	16: 14 female, 2 male. Older adults: 75-89. Vision, hearing, mobility impairments.	HOBBIT	Mixed method. Questionnaire interview, robot data. Home Interaction.	Keeping you occupied, did not feel safer. Attitude more negative over time, not emotionally attached. Should have round shapes, too large, different tray, too warm, fan-sound too high. Wished for more communication. Did not enhance independence. Easy to use, preferred buttons and touchscreen, forgot meaning of robots gestures.
(3) Bajones et al. (2018) Austria (Greece, Sweeden)	To present HOBBIT, a socially assistive service robot aiming at prolonged independent living of elderly people in their own homes	18: 16 female, 2 male. Older adults: 75-90. Vision, hearing, mobility impairments. Risk of falls	HOBBIT	Mixed method. Questionnaire interview, robot data. Private homes Interaction.	(3) Missing or non-working functionality.
(4) Bakracheva et al. (2020) Bulgaria	To analyze attitudes towards the functions the robots can perform to assist and improve daily life, preferences to robots’ control, appearance, and voice. Preferred activities, preference between and expectations to six versions of ROBCO.	32: 26 female, 5 male. With disabilities 25-80.	ROBCO	Mixed method. Questionnaire with open ended questions. Online survey with video. Club for disabled.	(1) Preffered the robot to be controlled by them (and relatives, other operators).(2) Age, gender differences to design.(3) Enhance independence. Assessment of upper limb not needed.(4) Preffered voice control, and remote control via internet. Mixed response on self-controlling.
(5) Bedaf et al. (2016) Netherlands (UK, France)	To develop the functionalities of Care-O-bot, discuss whether the presence of a service robot could be used to change behavior.	55: 36 female, 19 male. Receiving care, no cognitive decline Older adults: 62-95.	Care-O-bot	Qualitative. Interviews Video Universities.	Do not want the robot to override own decisions (autonomy). (3) Reminders, useful -could be irritating. Mixed on robot suggesting modification of users behavior. Safety, worried about dehydration if the robot refused to get a drink. Privacy, sharing data regarded as necessary. Robot seen as extension of care.
(6) Bedaf et al. (2018) Netherlands	To present how different stakeholders evaluate the role of a robotic assistant at home, what this means for future robot developments.	10: 7 female 3 male. Older adults: 62-93.	Care-O-bot	Mixed method. Questionnaire Interview. Simulated home. Interaction.	Positive experiences of the robot (3) Slow speed. Should do more complex task to contribute to independence. Functionalities are too limited. Wished for fetching and carrying objects, contacting others in case of emergency, opening door, cooking. Not supporting in showering, toileting and getting dressed. Should execute tasks based on users personal preferences. (4) Managed to use the GUI and control the robot by remote control. Struggeled to understand changing colors on the robot.
(7) Beer et al. (2015) USA	To explore the match between user and robot capability, and how this match may change as a function of task.	24: 12 female, 12 male. Half w/mobility issues. Older adults: 65-80.	PR2	Mixed method. Questionnaire Interview/TAL Video.	(3) Useful: Finding objects, dressing (older with mobility loss (wml), transfers, monitoring (wml), medication reminder. Speculations on robots capabilities. Distance-controlled robot could be beneficial.
(8) Beer et al. (2012) USA	Understand older adults’ preferences and perceived advantages/disadvantages of controlling a robot and assess older adults’ willingness to use control methods.	12: 6 female 6 male. Older adults: 68-79.	PR2	Mixed method. Questionnaire Interview Living lab. Interaction.	(3) Suggested robot to not give feedback after completing a task, half suggested visual feedback (e.g. blinking lights). Negative or mixed views on other persons controlling the robot. (4) Preferred voice command, not physical manipulation, somewhat positive to laser pointer.
(9) Beer et al. (2017) USA	To understand older adults’ first impressions (attitudes) of the assistive robot, if preferences and acceptance change between pre and post demonstration.	12: 6 female, 6 male. Older adults: 68-79.	PR2	Mixed method. Questionnaire Interview Living lab Interaction.	(3) Useful: Medication delivery, saves time and effort, too slow. Organizing objects. Preferred human in personal assistance, care, leisure- but robot in chores, information management, manipulating objects. (4) Exposure increased perceived usability and acceptance, more real life than video. Learning to use light switch: mixed responses.
(10) Bevilacqua et al. (2015) Italy	To critcally analyze the methodological choices done and investigate how to implement it in the second experimental loop of the Robot-Era project.	57: No info on gender. W/ health problems, mobility and cognitive deficits. Older adults 65+.	(Oro, Coro) Doro	Mixed method. Questionnaire Observation Living lab/ senior/ retirement home/local clinic. Interaction.	(3) Useful: Reminding service, Italy not Sweeden sample. (4) Services usable or excellent. Some expressed frustration, most enjoyment, and satisfaction. Difficulties with drop down menu.
(11) Bevilacqua et al. (2021) Italy	To explore the psychosocial determinants that lead to acceptability and willingness to interact with a service robot.	35: 22 female, 13 male. Older adults: mean age 73,8.	(Oro, Coro) Doro	Mixed method. Observation. Living lab. Interaction.	(1) Enjoyed using Doro, (one showed anxiety, two negative). (3) Garbage disposals, communication, receive warnings appreciated. Wanted more conversational communication and complex functionality.
(12) Brandl et al. (2016) Germany	To gain a deeper understanding of the distances that humans will accept between themselves and an approaching service robot.	30: 17 female, 13 male. 20-75	Care-O-bot	Mixed method. Distance data, Interview. Interaction.	Approaches increases closeness acceptance, appearing showing front or side does not affect acceptance. Closest acceptance was 0,5 cm in slow speed, participants preferred medium speed.
(13) Cavallo et al. (2014) Italy	To describe the Robot-Era project, which aims to develop, implement and demonstrate the feasibility, scientific/technical effectiveness and social/legal plausibility and acceptability of cooperating robots integrated with smart environments.	35: 22 female, 13 male. No cognitive decline. Older adults: 65-85.	(Oro, Coro) Doro	Mixed method. Data on time spent Questionnaire Interview. Homelab. Interaction.	Trusted the robot, safe. Aestethically pleasing, too big and bulky. Appreciated services, except shopping assistance. Enhance independence. Easy to use.
(14) Cavallo et al. (2018) Italy	To investigate the acceptance of the Robot-Era system, which provides robotic services to permit older people to remain in their homes.	45: No cognitive decline. Older adults: 65-86	(Oro, Coro) Doro	Quantitative. Questionnaire Living lab. Interaction.	Older users with a high education, more willing to interact than volunteers with a low level of education. Not judged dangerous. Feel safer due to warnings. Positive attitudes. Appreciated robot services. Understood functionalities very well. Will enhance independence. Walking assistance rated low as users did not have mobility issues. Shopping rated low.
(15) Chen et al. (2013) USA	To empower motor-impaired users to take full advantage of Assistive mobile manipulators (AMM) to perform a range of tasks, handle real-world variation found in homes and to inform the design of future AMMs.	Case study: 1 male. Disabled.	PR2	Mixed method. Questionnaire Statements. Home Interaction.	(3) Positive to tasks that involve manipulating objects remotely, such as tidying the house, answering the door, or fetching objects.
(16) Deutsch et al. (2019) Israel	Evaluating attitudes and emotional reactions towards different types of home robotic devices.	30: 20 female, 10 male. Healthy. Older adults: 67-90	PR2	Qualitative. Interview. Homes. Video.	Too big. Accepting of both human and object like looks. Positive to cleaning, a robot doing something meaningful, have purpose add to the quality of life. Negative to a robot moving around, a proactive robot, a robot with speech. Concerns about being too complicated
(17) DiNuovo et al. (2014) UK	To develop a more inclusive and usable user interface to facilitate interaction with the robots and to operate the novel service system (SUI/GUI).	82 Older adults.	(Oro, Coro) Doro	Mixed method. Questionnaire Observations. Living lab, Interaction.	Positive the overall experience. (3) Excellent: Shopping, Garbage disposal and communication. (4) Users with low technology experience had some difficulties using the tablet.
(18) DiNuovo et al. (2018) UK	To present the user-centred design approach of a multi-modal user interface (MMUI), which targets a more intuitive human way of interacting with a machine.	82: 51 female, 31 male. Healthy. Older adults 63-97.	(Oro, Coro) Doro	Mixed method. Questionnaire Interview Living lab Office room Interaction.	Positive acceptance by users. (4) Perceived useful, enjoyed a mix of SUI/GUI, scored high on usability for all tasks especially garbage disposal, communication and shopping. Multimodality, added value. Users with low technology experience had some difficulties using the tablet. Showed great tolerance to malfunctions.
(19) Eftring and Frennert (2016) Sweeden	To understand how the seniors used the robot and detect and prevent falls.	49 in lab/18 home, no info on gender or age. Risk of falls.	HOBBIT	Qualitative observations. Living lab Home Interaction.	Missed it after the intervention. Happy and friendly. Too big. Wanted it to manage thresholds and carpets. Keeps you occupied. Perceived it more as a toy/entertainment. Easy to use. Do not feel safer.
(20) Fischinger et al. (2016) Austria	To derive implications for improvement of the PT1 with respect to usability, acceptance, and affordability.	49: No info on gender. 70+. Vision, hearing, mobility impairments	HOBBIT	Mixed method. Questionnaire Interview. Living lab. Interaction.	Skeptical in the beginning but changed. Pick up object rated most useful. Concidered slow. Preferred voice command and multimodal control. Could imagine that HOBBIT took care of them. Liked the entertainment options. Perceived all tasks as rather easy and intuitively. Mixed opinions about its helpfulness.
(21) Foukarakis et al. (2017) Greece	To determine any specific technical difficulties that may occur in a real elderly persons’ home environment as opposed to the lab environment- by the use of FIRMA framework.	18: 6 female, 12 male. Healthy and mild cognitive disorders, mild Alzheimers. No info on age.	RAMCIP	Qualitative. Observation. Living lab. Interaction.	(4) Difficulties in hearing/understanding text to speech, preferred voice command, wished for easier user interface on Skype-communication mode. Preferred dark interface, small button sizes.
(22) Gerlowska et al. (2018) Poland	To verify users’ requirements implemented into particular HRI scenarios in terms of their perceived acceptability, usability, attractiveness. To establish the level of the societal impact of the robotic assistant perceived by the participants.	18: No info on gender. 10 healthy and 8 MCI. Older adults: 55-90.	RAMCIP	Quantitative Questionnaire Semi-controlled environment Interaction.	Positively influencing users’ mood and behavior, facilitating communication, would enhance the user’s security, benefit his/her quality of life, enhance everyday functioning positively influencing the functioning of the user. Efficiency rated low. Experienced the usability of the device as neutral.
(23) Grice et al. (2013) USA	If physical contact with the robotic system is acceptable.	2: Disabled persons 8: male able bodied mean age 26.	PR2 arm with tactile sensing skin.	Mixed method. Timing, Questionnaire interview, Home/no info. Interaction.	Able bodied: Felt safe. Disabled users: felt comfortable, safe. Felt soft, not too strong. Positive to further development.
(24) Grice and Kemp (2019) USA	To investigate the use of robotic body surrogates.	16: No info on gender or age. 15 novice users 1 expert.	Willow Garage PR2	Mixed method. Questionnaire Interviews Robot data. Users operate from home to lab.	(3) Would provide a meaningful improvement in their lives. (4) Easy to use.
(25) Hendrich et al. (2015) Germany	To outline an AAL architecture that targets integrated robotic services for elderly users spanning indoor, residential, and outdoor spaces.	70: Older adults. No info gender or age	Doro	Mixed method. Questionnaire Observation Age care facility. Interaction.	(2) Disliked the size and appearance. (4) Preferred the speech interface over the tablet.
(26) Kapusta et al. (2019) USA	Evaluation of a robotic system with a person with severe quadriplegia, in his home.	Case study, motor impaired person	Autobed combined with Willow Garage PR2	Qualitative. Free form feedback. Home, in own bed. Interaction.	(3) Liked that the robot autonomously reached to the task area in pulling up the blanket. (4) Removing a bed-blanket experieced as easy, wiping youghurt from the mouth more complicated, easier with practice.
(27) Koay et al. (2014) UK	To examine the relationship between user expectations, based on social roles, and proxemics preferences in an interaction with Care-O-bot 3	30: 11 female, 19 male. 19-25.	Care-O-bot	Quantitative, Questionnaire. Living lab. Interaction.	More comfortable with the robot over time. (3) Did not distinguish between in what distance they wanted the robot to approach for delivering different objects. (4) Struggled with recalling light colors and meaning of these.
(28) Körtner et al. (2014) Austria	Developing a highly acceptable and affordable socially assistive robot supporting older adults in staying independently at home.	49: 35 female, 14 male. Older adults: 70-88, Vision, hearing, mobility impairments.	HOBBIT	Mixed methods. Questionnaire Interview Interaction.	(2) Enjoyed face and voice. Right size or smaller. (3) Experienced picking up objects, using robot as walking support, getting up of chair or after fall as very useful. Speed too slow. Can imagine robot able to take care of them. (4) Experienced controlling the robot as easy, liked both voice command and touch screen, few liked control by gestures.
(29) Lammer et al. (2014) Austria, Sweeden, Greece	To investigate the difference in the user perception of the HOBBIT robot with reciprocity fostering and normal dialogue behavior.	49: (25 controls), no info on gender. Older adults: 70+. Vision, hearing, mobility impairments and heallthy.	HOBBIT	Mixed method. Questionnaire observation. Laboratories Interaction.	(4) Positive or neutral during the tasks of the robot when bringing an object with failure. Enjoyed being surprised and being asked for help by the robot. (5) Majority of participants wanted to be called by their real names by the robot. (4) Participants helped the robot when it needed help.
(30) Lee et al. (2016) USA	To analyze how older adults make sense of assistive robots based on their everyday life and aging experiences.	7: 6 female, 1 male. Healthy, 1 person w disability. Older adults: 57-72.	Care-O-bot	Qualitative Interview Video Homes.	(1) Experienced the robot as being helpful for older adults, worried about privacy. (3) Described as “wonderful,” “great,” or “work(ing) great”, did not see themselves as the targeted audience, did not identify with the stereo-type user.
(31) Lee and Riek (2018) USA	To explore alternative ways to frame aging in assistive robotics, investigate alternative robot designs for aging based on our participants’ interpretation of aging.	9: 5 female, 4 male. Older adults: 65-85.	Care-O-bot, (Riba)	Qualitative Collaborative Interview. Cafes, residential centers, home. Video.	(1) Expressed concerns about robots monitoring all the time (3) Suggested robots with sympathy and robots that would enhance resilience.
(32) Mejia (2015) Japan	Demonstrate that interaction decisions, in terms of how an avatar and a GUI are displayed in the same screen, affect user’s impressions towards the robotics system.	24: 16 female 8 male. Older adults. No info on age.	HomeMate	Qualitative Interview observation, reaction cards. Age care facility. Interaction.	(2) Enthusiastic about interacting with the robot for longer periods of time when avatar in display with smaller GUI in categories. (3) Users was friendlier towards the robot when the interface was dominated by an avatar. Found it complicated to remember and learn the icons.
(33) Park et al. (2020) USA	To introduce and evaluate a proof-of-concept of meal-assistance system using a general-purpose mobile manipulator, a PR2 robot.	18: 9 able bodied:1 female, 8 male. 22-29 8 Motor impaired: 4 male, 4 female. 23-72.	PR2	Mixed method. Questionnaire Observation Lab, and 1 users home. Interaction.	Experienced the system safe, successful and comfortable to use. (3) One participant wished for faster feeding.
(34) Piasek and Tobis (2018) Poland	To develop and evaluate the impact of new solutions in service robotics for assisted living environments to support older people affected by MCI.	5: 4 female, 1 male. Mild cognitive decline. Older adults 66-83.	TIAGo	Quantitative Questionnaire Homes. Interaction.	(5) Experienced the robot as ”definitely not scary” positive feelings during the interaction. Described as useful and likable.
(35) Prakash et al. (2014) USA	Obtaining opinions about the appearance of the robot to identify the categories of considerations that are important to older adults.	12: 6 female, 6 male Older adults: 68-79.	PR2	Qualitative. Interviews. Homes. Interaction.	(2) Wanted their robot to look like a human or have human characteristics. They preffered non-scary, smaller sizes that would fit the home. Liked the human likeness and maneuverability. Disliked size, weight, laser scanner in the front. Most participants agreed on importance of appearance, two stated that functionalities were more important.
(36) Prakash et al. (2013) USA	To gain insights into factors that influence older adults’ attitudes toward a mobile manipulator robot for supporting medication management.	12: 6 female, 6 male. Older adults: 68-79.	PR2	Mixed method. Questionnaire Interview. Living lab. Interaction.	(3) Robots speed 10cm/s comfortable. Mixed on direction of approach. Did not intuitively understand when and how the gripper would release the medication. Willing to use the RFID necklace and to label the medication bottles, if that made the robot able to assist. Positive to reminder of medications intake, mixed on interruptions. Preferred the whole box (reliability) rather than a single pill.
(37) Pripfl et al. (2016a) Austria	Develop a low-cost robot that assists in daily living of senior citizens in private homes. Reduce the risk of falling, detect falls and handle emergencies.	7: 6 female, 1 male. Vision, hearing, mobility impairments Older adults: 75-88.	HOBBIT	Mixed method. Questionnaire Interview Homes. Interaction.	No enhanced feeling of safety at home. No perception of increased own independence (3) Appreciated picking up objects from the floor, transporting objects, emergency recognition, fitness supervision and giving reminders. (4) Errors in the actions of the robot led to frustration.
(38) Pripfl et al. (2016b) participantsAustria	To investigate how seniors experienced the interaction with the robot at home, in terms of usability, utility, support of their independent living, and their feelings of safety.	18/16: 14 female, 4 male. Older adults: 75-89.	HOBBIT	Mixed method Questionnaire Interview Homes. Interaction.	(3) Appreciated the functions of picking up objects from the floor, transporting objects, emergency recognition, fitness program, and giving reminders. Experienced the speed as too slow. (4) Prototype was intuitive to handle, but errors in the actions of the robot led to frustration.
(39) Rosenthal and Krämer (2015) Germany	To explore participants’ evaluations of human-like robots, their attitudes and emotional reactions.	38: 16: 8 female, 8 male. 27-68 22: 5-11.	AR	Qualitative, interviews. Homes/school/university. Video.	Experienced the robot as useful, positive to owning it. Some people liked the more abstract appearance with not too many human-like features. Positive to its movements. Felt safe. Household chores suitable task for robots.
(40) Saunders et al. (2016) UK	Describe an approach to service robot personalization designed to be used by carers, relatives, and elderly persons themselves.	20: 16 female, 4 male. 20-67.	Care-O-bot	Quantitative, Questionnaire Likert scale. Living lab. Interaction.	(3) Useful to teach the robot tasks, comfortable with the robot informing of a problem in the house. (4) They were able teach the robot tasks. They experienced the interface easy to use. Older aged had more difficulties in using the system.
(41) Smarr et al. (2012) USA	To determine whether older adults are open to accept robot assistance in the home, identify the tasks that older adults would accept robot assistance with in the home and categorize those tasks.	21: 15 female, 6 male. Healthy. Older adults: 65-93	PR2	Mixed method, Questionnaire Interviews No info on context. Video.	Generally willing to accept robots. (4) Willing to accept assistance from a robot in tasks within the home. Preferred assistance from a robot over a human for performing IADLs and EADLs. Less receptive to robot assistance in performing ADLs (e.g., eating/feeding self, getting dressed, shaving, walking).
(42) Smarr et al. (2014) USA	To discover the American older adults acceptance of robot assistance in the home, which actions they want robots to perform and objects they want the robots to manipulate.	21: 15 female, 6 male. Good health. Older adults: 65-93.	PR2	Mixed method Questionnaire interviews No info on context. Video.	Generally willing to accept robots. (3) Willing to accept assistance from a robot in tasks within the home. Preferred assistance from a robot over a human for performing IADLs and EADLs. Less receptive to robot assistance in performing ADLs (e.g., eating/feeding self, getting dressed, shaving, walking).
(43) Vincze et al. (2014) Austria	To test the difference in the user perception of the HOBBIT robot with a reciprocity fostering behavior and normal dialogue behavior.	49: Older adults. No info on gender or age.	HOBBIT	Mixed method Questionnaireobservation No info on context. Interaction.	Compliance and emotional responses were positive or neutral during every step of the robot task. (3) Indicate that the robot could be very helpful.(4) Easy to use.
(44) Wang et al. (2016) UK	Describe and evaluate the speech interface	47: 26 female, 21 male. Older adults: 65-79.	Doro	Quantitative Questionnaire. Testsite. Interaction.	(3) More satisfied with the SUI experience for the four services: shopping, reminding, garbage collection, and indoor escort.

Abbreviations: The table show the included papers in the Scoping Review. The aim of the papers, information provided about the participants, method and context and the participants experience of the robot after interaction or video of the robots. Both quantitative and qualitative papers are included. GUI: Graphical interface. SUI = Speech user interface. Cm/s: centimeter per second. EADL= Enhanced activities of daily living, include participation in social and enriching activities. IADL= Instrumental activities of daily living, include the ability to e.g., use the telephone, shop, prepare food, do the housekeeping and laundry, manage medications and finances, and use transportation. MCI= Mild cognitive impairments. AAL= Ambient assisted living. In papers including more than one robot, data have been extracted for the robots able to assist physically. In paper including both caretakers, health professionals and similar- data have been extracted for participants receiving assistance from a robot.

There are 33 main authors for the 44 papers. Several of these papers refer to the same research project and robot (Table 4).

Table 4. Included papers connected to the same project.

Authors	Robot	Project
Bedaf et al. (2016/2018), Koay et al. (2014), Saunders et al. (2016)	Care-O-bot	ACCOMPANY: Acceptable Robotic Companions for the Aging Years
Piasek & Wieczorowska-Tobis (2018)	TIAGo	EnrichMe
Bajones et al. (2018, 2020), Eftring & Frennert (2016), Fischinger et al. (2016), KÖrtner et al. (2014), Lammer et al. (2014), Vincze et al. (2014)	HOBBIT	HOBBIT
Foukarakis et al. (2017), Gerlowska et al. (2018)	RAMCIP	RAMCIP: Robotic Assistant for MCI Patients at home Robotic Assistant for MCI Patients
Bevilacqua et al. (2015, 2021), Cavallo et al. (2014, 2018), DiNuovo et al. (2014/2018), Hendrich et al. (2015), Wang et al. (2015)	Doro	Robot-Era: Implementation and integration of advanced Robotic systems and intelligent Environments in real scenarios for the ageing population
Chen et al. (2013), Grice et al. (2013)	PR2	Robots for Humanity
Bakracheva et al. (2020)	ROBCO	Tele-controlled service robots to improve the quality of life of adults and the disabled
Brandl et al. (2016)	Care-O-bot	Tech4p/MeRoSy: Mensch Roboter Evolution

Abbreviations: MCI: Mild Cognitive Impairment.

In 33 papers, the participants interact with a robot. In ten studies, the participants are shown videos of a robot in action, and in one study the participants operate the robot remotely (Grice & Kemp, 2019). In 33 of the papers, it is evident that the investigation primarily focuses on, or include older adults. A few studies include younger participants (Arnold & Scheutz, 2018; Bakracheva et al., 2020; Brandl et al., 2016; Chen et al., 2013; Grice et al., 2013; Hendrich et al., 2015; Park et al., 2020; Rosenthal von der Pütten & Krämer, 2015). The interaction with robots has been performed in real homes (including age care facilities), simulated homes (living-labs) and other settings like universities and laboratories.

The included robots and how they have been used to assist physically in ADL

The robots studied in the papers (Table 5) are all mobile robots, mounted on a wheeled base. They exhibit various forms of torsos, and heads. The Care-O-bot`s head is affixed to the torso, while the heads of the RAMCIP and HomeMates’ are animations on a tablet. HOBBIT, PR2, Doro, TIAGo and Artificial Robot (AR) have more human-like heads. The robots are equipped with either one or two arms featuring grippers, enabling them to assist in ADL activities. The robots have the ability to either navigate automatically or under the control of an operator or user. Table 5 illustrates how these different humanoid robots have been employed to assist people with ADL, encompassing tasks such as, fetching and retrieving objects, garbage disposal, tidying up and assisting with laundry. For more comprehensive technological descriptions the papers and in some cases the projects, distributor’s websites can be consulted. Table 6 outline the status of the robots,’ the author, distributor, or website.

Table 5. Robots and functions outlined in the papers.

Robots	AR	Care-O-Bot	Doro	HOBBIT	HomeMate	PR2	RAMCIP	ROBCO	TIAGo
Bring objects		x	x	x	x	x	x
Pick up objects				x	x	x	x
Deliver medication						x
Feeding						x
Shaving						x
Shopping assistance			x
Garbage disposal			x
Dressing						x
Pull up blanket						x
Folding towels						x
Turn on light						x	x
Entertainment				x	x				x
Exercise				x					x
Monitor/alarm		x		x			x	x	x
Videophone			x	x	x		x		x
Household chores e.g., laundry, clean up table	x		x			x	x		x
Reminders		x	x	x		x
Walking escort			x
Limited information	x			x	x			x

Images with permission from: AR: (Jouhou-System-Kougaku-Laboratory, 2022), Care-O-bot: (Fraunhofer-IPA, 2010), Doro (Hendrich et al., 2015), HOBBIT: (M.Vincze), HomeMate (Sukhan Lee, Sungkyunkwan University), Pr2: (©John Greenleigh/Flipside Studios), RAMCIP: (Kostavelis, 2018), ROBCO (Bakracheva et al., 2020), TIAGo: (© PAL Robotics 2022).

Table 6. Robot status.

Robot	Information found	Status (websites and authors)	Commercialized
AR	2009	Continuing development on other versions	–
Care–O–bot	1998	Version 4 further developed	X
Doro	2014	No info	–
HOBBIT	2014	No further development	–
HomeMate	2012	Version 4 further developed	–
PR2	1998	No further development	–
RAMCIP	2015	Further developed	–
ROBCO	2009	No information	–
TIAGo	2016	Further developed	X

Users’ experiences of domestic humanoid robots assisting in ADL

Four categories of experiences related to use of the robots were identified through the included papers: (1) the robots’ ability to assist with ADL, (2) the robots` appearance, (3) users’ sense of safety and privacy, and (4) ease of use. Results are provided with a numerical representation when provided in the papers.

The robots’ ability to assist with ADL

Ar

Participants expressed positive reactions after watching videos of the robot AR doing laundry. They considered it useful and were interested in having it. They scored the robot high on likability and considered household chores as suitable for the robot (Rosenthal von der Pütten & Krämer, 2015).

Care-O-bot

All participants found it useful to teach the robot how to perform new behaviors such as prepare a meal, and most (n = 17/19) were comfortable with the robot notifying both them and others of a problem in the home (Saunders et al., 2016).

Participants had mixed responses to whether the robot should prompt healthy behavior. While they found the reminder function useful, some considered it potentially annoying (Bedaf et al., 2016). Users suggested that the robot should possess the ability to cook, open doors, fetch and carry objects and contact help in an emergency. However they did not want the robot to support in showering, toileting or getting dressed (Bedaf et al., 2018). In a study by Lee et al. (2016) the participants were found to be overly favorable toward robots (n = 5/7). However, they did not see themselves as the target audience, and associated robot assistance more with disability than age.

Doro (Oro, Coro)

The Robot-Era project entails the collaboration of three robots, each with distinct functionalities. Of the three only Doro falls within our inclusion criteria.

Most users appreciated the robots’ services. The assistance in disposing garbage was particularly mentioned. Participants appreciated communicating with the robot and stated that the robot would ease the burden on family members who provided care (Bevilacqua et al., 2015, 2021; DiNuovo et al., 2014, 2018). They wanted to communicate more with the robot and wished for more complex functionalities in Doro (Bevilacqua et al., 2021). Assistance in grocery delivery was not desired by the participants in Cavallo et al. (2014) study, as they found shopping an enjoyable task that improved their quality of life. The participants that did not have mobility issues did not see the need for the robot assisting with walking (Cavallo et al., 2018).

HOBBIT

Participants were positive or indifferent to the tasks they tried out with the robot (Lammer et al., 2014; Vincze et al., 2014). They expressed appreciation for various functions including carrying objects, emergency recognition, fitness programs, reminders (Pripfl, et al., 2016b), walking support, assistance rising from a chair or from a fall (Körtner et al., 2014), and notably picking up dropped objects (Fischinger et al., 2016; Körtner et al., 2014; Pripfl, 2016a; Pripfl, Körtner, Batko-Klein, Hebesberger, Weninger, Gisinger, Frennert, et al., 2016b). However, some users found the robot too slow (Fischinger et al., 2016; Körtner et al., 2014; Pripfl, et al., 2016b). While a subset of participants perceived the robot more as a toy (Bajones et al., 2020; Eftring & Frennert, 2016) and not contributing significantly to their independence (Pripfl, 2016a; Pripfl, Körtner, Batko-Klein, Hebesberger, Weninger, Gisinger, Frennert, et al., 2016b), they believed it could be helpful in the future (Bajones et al., 2020; Fischinger et al., 2016; Körtner et al., 2014). In Fischinger et al. (2016) the participants (65%) stated that they could envision having the robot take care of them. However, 45% were skeptical, expressing the preference for a human helper.

HomeMate

The older adults in the study of HomeMate, communicated with the robot through the tablet and provided feedback on the tablets design. Participants found the robot friendlier and were willing to interact with it for longer periods of time, when the interface had a human avatar. Mejia (2015) found that combining avatars and graphical user interfaces, can improve users’ perceptions of robot’s capabilities.

PR2

Beer et al. (2017) found that participants (n = 9/12) were positive to the robot delivering medication. They expressed that it could save time and effort, particularly favoring the robot delivering medication bottles over single pills. The robot’s proficiency in finding objects, clean the house and assistance with transfers were deemed most useful. Participants appreciated the idea of teaching the robot new tasks and liked the way that the robot organized objects. While they preferred a human to assist with care, leisure, and personal assistance, they leaned toward a robot to do chores, information management and object manipulation (Beer et al., 2017; Smarr et al., 2014). A few participants preferred having the robot initiate actions and to speak (9/25 quotes), but most participants did not (16/25 quotes) (Deutsch et al., 2019). Users expressed a desire for the robot in providing feeding assistance and to handle tasks such as answering the door and assist with dressing, including putting on socks and fastening bras (Chen et al., 2013; Park et al., 2020).

RAMCIP

Participants in Gerlowska et al. (2018) found that the robot would enhance their quality of life, improve everyday functioning, and lighten the burden on caregivers. However, the efficiency of the RAMCIP was rated as low, attributed to the current functionality and participants’ somewhat unrealistic expectations (Gerlowska et al., 2018).

ROBCO

Upon watching videos of ROBCO, all (n = 32) participants expressed the belief that the robot would enhance their independence. They preferred the robot`s assistance in carrying heavy objects, bring water and food, issue warnings about possible deterioration, aiding in shopping and supporting the visually impaired. While they found the rehabilitation function useful, the ability to assess upper limb function was not deemed as valuable (Bakracheva et al., 2020).

TIAGo

TIAGo received positive ratings for being useful and likeable in tasks like picking up objects, entertainment and guiding in physical exercises. Although there was, a small difference between users with and without prior experience with technologies, overall, users reported positive feelings during the 10-week interaction with TIAGo (Piasek & Wieczorowska-Tobis, 2018).

Experience of the robot’s appearance

Some of the participants stated that they found the Robot-Era robots too big and bulky (Cavallo et al., 2014; Hendrich et al., 2015). This was particularly expressed for the PR2 robot (Deutsch et al., 2019; Prakash et al., 2014). Participants disliked the laser scanner in the front of PR2. They did, however, agree that the robot’s functionalities were more important than its appearance.

There were mixed reports on Doro’s appearance (Cavallo et al., 2014; Hendrich et al., 2015)

Participants enjoyed the HOBBIT’s voice (88%) and face (69%) (Körtner et al., 2014) and missed the robot after the trial ended (Eftring & Frennert, 2016). Participants reacted to the HOBBIT as being too warm, and the sound from its fan being too loud, but it was also considered cute (Bajones et al., 2020). Participants stated that they wanted their robot to look human and have human characteristics. Otherwise it should look elegant, beautiful and non-intrusive (Prakash et al., 2014).

Feelings on safety and privacy

None of the participants in the papers reported feeling scared or unsafe around the robots. However, participants interacting with HOBBIT did not think that its presence made them feel safer, especially in scenarios like falling (Eftring & Frennert, 2016; Pripfl, 2016a; Pripfl, Körtner, Batko-Klein, Hebesberger, Weninger, Gisinger, Frennert, et al., 2016b), and users (62,5%) reported a higher degree of anxiety and insecurity over time with HOBBIT (Bajones et al., 2020). Some participants expressed concerns that the Care-O-bot might refuse to perform tasks, such as bringing a drink, potentially leading to dehydration. Piasek and Wieczorowska-Tobis (2018) found that participants felt calm during their interactions with TIAGo. Those interacting with RAMCIP expressed that the robot would enhance their sense of safety (Gerlowska et al., 2018). Cavallo et al. (2018), found that participants felt safer due to robot warnings. Moreover, participants believed they would feel at a lower risk of falling, since the robot could both bring them what they needed and function as walking support (Doro). Participants watching the AR robot evaluated it as safe (Rosenthal von der Pütten & Krämer, 2015). Participants reported feeling safe in close encounters with PR2, such as feeding and shaving (Grice et al., 2013; Park et al., 2020).

Concerning privacy, participants reported mixed experiences. Cavallo et al. (2018) participants did not consider the robots intrusive. Bedaf et al. (2016) found that the participants saw data sharing as necessary when receiving assistance from Care-O-bot. However, they did not want the robot to override their decisions and emphasized that users would have to agree to assistance. Others worried about their privacy when being monitored by the robot (Lee & Riek, 2018; Lee et al., 2016).

Perceived ease of use

The papers detail participants’ experiences using the robots with different interfaces. HOBBIT received positive feedback and was described as easy or rather easy to use (Bajones et al., 2018; Eftring & Frennert, 2016; Fischinger et al., 2016; Pripfl, et al., 2014). However, users had various experiences in controlling the robot. Some (49%) preferred voice commands, despite occasional misunderstandings (Körtner et al., 2014). Hearing-impaired participants preferred the touchscreen and separate buttons. Users disliked the HOBBIT communicating through gestures, as they often forgot what the gestures meant (Bajones et al., 2020; Körtner et al., 2014). Some expressed frustration with errors, technical difficulties, or the robot’s failure to understand voice commands. The participants perceived that usability increased after interacting with the PR2 (Beer et al., 2017; Smarr et al., 2014).

For ROBCO, voice control, and remote control through internet was rated as the most appropriate. Controlling the robot’s arm and grippers, received mixed response with, 55% considering it easy, and 44% difficult. The majority preferred to self-control but would also accept remote control by relatives or health professionals (Bakracheva et al., 2020). Regarding PR2, participants who controlled the robot, found some tasks easy to perform, and that harder tasks became easier with practice (Grice & Kemp, 2019; Kapusta et al., 2019; Park et al., 2020). Beer et al. (2012) found that participants preferred voice command (83%), but others did not mind using a joystick (Likert scale 3,25/5) or a laser pointer to control the PR2 (Likert scale 3,25/5). There was disagreement on whether others should control the robot remotely.

Most participants enjoyed using Doro according to Bevilacqua et al (2015, 2021). Users without technology experience faced difficulties using the control tablet (DiNuovo et al., 2014). The graphical user interface was praised despite occasionally technology failures (Bevilacqua et al., 2015). Preferences for interface combinations varied; some users preferred a mix of speech and graphic user interfaces, while others leaned toward the speech user interface (DiNuovo et al., 2014; Hendrich et al., 2015; Wang et al., 2016). Some of the users had difficulties with the dropdown menus on the tablet. Users needed support to control the robot and frustration arose during certain tasks (Bevilacqua et al., 2015). Initial trouble with the speech interface improved with training (DiNuovo et al., 2014, 2018).

Care-O-bots use of different light colors for communication, posed challenges in recalling their meaning (Bedaf et al., 2018; Koay et al., 2014). Younger participants found the interface on the robot easier to use than the older participants (Saunders et al., 2016). Users had mixed feelings on whether someone else should set up the robot. In Bedaf et al. (2018), all the participants managed to complete the selected tasks and felt a sense of collaboration with the robot.

Participants in Mejia (2015) struggled to remember and learn icons on the graphical user interface (tablet) on HomeMate.

Discussion

The robots identified in this review have several software capabilities, like virtual communication and entertainment, but can only provide a limited number of physical assistance services within the ICF-framework domains of self-care, mobility, and domestic life. Within the mobility-related activities, we found robots that could bring the users an object, a drink or a medication, and pick up something that had fallen to the floor. Some robots could assist in transfer and support users when walking. In the domain of domestic life, some robots could do household chores. One robot could load and turn on a washing machine, another could assist in shopping and put garbage into a bin. One robot could fold towels. Another could assist users with self-care like covering them with a blanket, feeding them, wiping the mouth and shaving. These activities have all been highlighted within ICF-domains that best describe the spectrum of prototypical problems in functioning and health in people after suffering from stroke, spinal cord injury, cerebral palsy, and the elderly (Bedaf et al., 2014; ICF-Research-Branch, 2017). It is therefore likely that users with physical disabilities could benefit from the assistance of robots. Hence, developers of assistive robots could apply the ICF framework to align the robots’ capabilities with users’ limitations.

Older adults favored robots that could perform tasks like retrieving and removing items and disposing of garbage. Some expressed that the robot could increase their independence and improve their quality of life. This is in line with research (Venkatesh et al., 2003), which shows that performance expectancy or perceived usefulness is one of the most important factors in user acceptance of technology. However, as in Wu et al. (2016) study, several participants mentioned that the robots assisted them when it was unnecessary. This is because in many of the included papers, older adults were recruited on the basis of their age, not because they had needs that the robots could fulfill. This required the older adults to imagine that they would experience decline and answer questions about perceived future needs. Some participants may have found this difficult.

Wu et al. (2016) also described that the reluctance to use robot technology could be a form of passive resistance, explained by a fear of adopting or accepting the use of assistive technology would lead to stigmatization. Although fear or anxiety itself is not a direct determent of intention to use new technology in the UTAUT (Venkatesh et al., 2003) or the Almere model (Heerink et al., 2010), both age and experience were found to moderate users’ performance expectancy, or ease of use, and thereby intention to use. This implies that with practice and familiarity with a technology, users may find the systems easier to use and start using it more. This was confirmed by Beer et al. (2017).

We found no examples of papers that presented cases where users had purchased or otherwise fully implemented a robot in their home, and only a few studies were performed in users’ homes. This could be attributed to the insufficient technical readiness of these robots. As healthcare technologies must adhere to ISO standards for certification (International Organization for Standardization 2022), assistive robots face the critical requirement of operating reliably in diverse conditions. Ensuring robustness in navigation, object recognition, and overall system stability is essential for establishing user trust and acceptance. Technical challenges such as sensor errors or system failures can pose obstacles to the successful deployment of these robots (Papadopoulos et al., 2020) but also jeopardize user safety, potentially giving rise to adverse events.

Many daily tasks in a home environment involve a level of complexity that current assistive robots may struggle to handle. Tasks such as navigating through cluttered spaces, object manipulation, and responding to unstructured environments pose significant technical challenges.

This may imply that assistive robots need to be highly adaptable to diverse home environments and user needs. Achieving this adaptability in a cost-effective manner is a technical hurdle, requiring advanced sensors, machine learning algorithms, and sophisticated robotic hardware.

Participants frequently commented that the robots should be able to perform more complex tasks to support their independence. Users’ expectations of the performance and usefulness of a technological system were closely linked to the intention of use, as proposed by Venkatesh et al. (2003).

The participants mentioned the importance of personalizing the robots. The literature has addressed this finding (Andtfolk et al., 2021; Papadopoulos et al., 2020; Pino et al., 2015). Personalization has shown to affect users’ acceptance of robots, which will determine their willingness to use the robots (Pino et al., 2015). The robot must not perform activities that the users enjoy doing and that keep them active. The robot must, however, assist with activities that users cannot perform, or that deprive them of energy they would rather spend on other activities. The personalization of a robot’s features will of course be both expensive and time consuming and require access to technical maintenance since users’ requirements are bound to change with their health status.

Based on our results we advocate for a multidisciplinary approach to developing future robots for physical assistance, employing co-creation methods, that involve all stakeholders including developers, potential users and caregivers. The ICF framework can serve as a foundation for a practical guideline that could identify specific domains relevant in the context of assistive robotics. Such a guideline could assist in translating the concepts from the ICF framework into practical requirements for assistive robots. For example, if mobility is a key domain, consider how the robot can assist with mobility challenges, such as helping users move from one location to another. A guideline could further suggest how to assess how the robot can adapt to different environmental contexts and individual user characteristics to enhance its effectiveness and usability. And also, how to consider the diverse preferences and requirements of users with varying abilities and disabilities and ensure that the development and deployment of robots align with ethical standards outlined in the ICF framework.

Four papers discuss the topic of whether services could be delivered by a robot or by a human (Beer et al., 2017; Fischinger et al., 2016; Smarr et al., 2012, 2014). Some participants did not want robots to assist them with anything; others stated that they would use a robot for some practical tasks or to do chores but would prefer a human for personal care tasks. These findings could be related to the lack of trust in the robot’s ability to assist with personal care. Trust in the robot is essential to accept its delivery of services. The robots’ characteristics and performance have been found to be the most influential factors in the development of trust (Hancock et al., 2011). Grice and Kemp (2019) and Kapusta et al. (2019) show that a robot can safely provide care assistance, and that users accept and appreciate this assistance. Hence, robot developers should recognize that future users will experience robots differently, and that certain assistance tasks that are accepted by some users might be rejected by others.

The integration of humanoid assistive robots in healthcare brings up ethical and societal concerns. For example, these robots may gather and process sensitive patient data, emphasizing the need for robust security measures to safeguard patient privacy (Boada et al., 2021). In addition to meeting ISO standards, clear guidelines for obtaining patient consent regarding data collection and storage are essential to address privacy issues (Battistuzzi et al., 2018; International Organization for Standardization, 2022).

Furthermore, defining the extent of humanoid robots’ involvement in decision-making processes without infringing on patient autonomy is crucial. As found in Bedaf et al. (2016) the robot must not override the users’ decisions, and the user must agree to have a robotic helper (Bedaf et al., 2016). Ethical design, emphasizing empathy and cultural sensitivity, is also vital for positive human-robot interactions (Battistuzzi et al., 2018, Boada et al., 2021). Accessibility for all individuals, regardless of socioeconomic status or geographical location, is essential to avoid healthcare disparities driven by technology. As concerns about being able to afford the robot was mentioned by participants in the review, reasonable implementation costs are crucial to prevent a divide between institutions that can and cannot afford humanoid technology. To optimize the benefits of humanoid robots and prevent misuse, healthcare professionals and patients require proper training. In addition, strategies for retraining and providing alternative employment opportunities need consideration in situations where robots may replace human caregivers. Designing humanoid robots sensitive to diverse cultural norms is crucial for acceptance across different communities (Battistuzzi et al., 2018).

The results of this review indicate that these robots need further development before they will be ready for implementation. Hence, our review may provide valuable information for developers about users´ experiences of humanoid robots assisting in ADL, and developments needed in provision of physical assistive robotic functions within the ICF categories of self-care, mobility and domestic life.

Considering each user’s needs and supporting them with information to set realistic expectations will be a future concern for developers and decision makers in health service provision.

Limitations

While we demonstrated transparency by utilizing the Joanna Briggs framework, we recognize that our research team could have benefitted from incorporating technical expertise. Despite conducting a thorough search across health-related and technical databases and journals, it is improbable that we identified all studies within our scope. We anticipate that certain studies, particularly those detailing technical features and possibly incorporating user tests not explicitly mentioned in titles or abstracts, might have been unintentionally excluded. We did not perform internet or forward citation searches in this review. Limitations of our study include the exclusive review of English papers, potentially overlooking relevant publications in other languages. Recognizing that pertinent studies might have been conducted in regions such as Asia, a more diverse representation of nationalities in our research team could have enhanced the comprehensiveness of our analysis.

Many of the studies describe the same projects. A limitation of this study might then be that some results are overlapping, this however relates mainly to qualitative statements. Table 4 was developed for transparency regarding overlapping papers and projects.

Conclusion

In this Scoping Review of humanoid robots assisting in daily activities, we analyzed 44 papers focusing on nine robots able to assist users with physical disabilities. These robots can perform tasks such as finding objects and doing household chores, with one capable of assisting in care tasks. Despite the robots’ diverse capabilities, the readiness for physical assistance falls short for widespread implementation. While most participants expressed positivity toward the robots and their helpful services, concerns included dissatisfaction with size and speed, and some faced challenges with interfaces and voice control. The size of the humanoid robots could be seen as intimidating or overwhelming for certain individuals, especially when space is limited, making their presence cumbersome. Dissatisfaction may arise if users have high expectations for the speed of services, as any perceived delay can lead to dissatisfaction. Challenges with the interface, such as difficulty understanding commands or misinterpretations, can contribute to frustration and dissatisfaction, ultimately leading to nonadopting of the technology.

Participants desired robots with enhanced functionalities and personalization options. Although many found the services interesting and useful, some did not perceive them as necessary. This study highlights the intricate landscape of assistive robots, revealing both their capabilities and limitations. These insights bear substantial implications for the future of healthcare and the development of assistive technology. The study underscores the imperative for refining physical assistance capabilities, aligning development of robots with the ICF framework, and addressing user expectations, trust factors, technical challenges, societal attitudes, ethical considerations and context. We suggest that future research utilize the ICF framework and adopt a user-centric and multidisciplinary approach to ensuring that assistive robots effectively meet the diverse and evolving needs of users, contributing to an enhanced quality of life for individuals with physical disabilities.

Programming robots to navigate ethical dilemmas in alignment with human values is a complex challenge requiring careful consideration. Assessing the potential impact of humanoid robots on relationships and interpersonal dynamics is vital for understanding broader societal implications (Boada et al., 2021). To harness the benefits of humanoid assistive robots responsibly, it is crucial to address these ethical and societal considerations in healthcare integration. Continuous dialogue among healthcare professionals, technologists, ethicists, and the general public is essential to navigate these complex issues.

Acknowledgements

The authors would like to thank Ellen Sejersted for highly appreciated contribution to this work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The study was funded by the University of Agder. However, the funder had no role in the study or writing of the manuscript.