Recommendations for the Development of a Robotic Drinking and Eating Aid - An Ethnographic Study

Abstract

Being able to live independently and self-determined in one’s own home is a crucial factor or human dignity and preservation of self-worth. For people with severe physical impairments who cannot use their limbs for every day tasks, living in their own home is only possible with assistance from others. The inability to move arms and hands makes it hard to take care of oneself, e.g. drinking and eating independently. In this paper, we investigate how 15 participants with disabilities consume food and drinks. We report on interviews, participatory observations, and analyzed the aids they currently use. Based on our findings, we derive a set of recommendations that supports researchers and practitioners in designing future robotic drinking and eating aids for people with disabilities.

Full text

Recommendations for the Development of a
Robotic Drinking and Eating Aid - An
Ethnographic Study
Max Pascher1,2[0000−0002−6847−0696], Annalies Baumeister3, Stefan
Schneegass2[0000−0002−0132−4934], Barbara Klein3, and Jens Gerken1
1Westphalian University of Applied Sciences
2University of Duisburg-Essen
3Frankfurt University of Applied Sciences
(a) Filling a glass of water (b) Grasping a glass to drink
Fig. 1: Robotic arms can support users with motor impairments in their everyday
drinking and eating task. We explore how such systems should be designed to
provide a benefit to the users and support them in living a self-determined life.
Abstract. Being able to live independently and self-determined in one’s
own home is a crucial factor for human dignity and preservation of self-
worth. For people with severe physical impairments who cannot use their
limbs for every day tasks, living in their own home is only possible with
assistance from others. The inability to move arms and hands makes it
hard to take care of oneself, e. g. drinking and eating independently. In
this paper, we investigate how 15 participants with disabilities consume
food and drinks. We report on interviews, participatory observations,
and analyzed the aids they currently use. Based on our findings, we
derive a set of recommendations that supports researchers and practi-
tioners in designing future robotic drinking and eating aids for people
with disabilities.
Keywords: Assisted Living Technologies ·Human-Centered Comput-
ing ·Meal Assistance ·Participation Design ·People with Disabilities ·
Robot Assistive Drinking ·Robot Assistive Feeding ·User Acceptance ·
User-Centered Design ·User Participation.

2 M. Pascher et al.
1 Introduction
At the end of 2019, 7.9 million people classed as severely disabled were living
in Germany [47]. With over 58% of these cases being attributed to physical
disabilities, motor impairments affected an total of 4.6 million people; 11.2%
of which are suffering from impaired functionality to a complete loss of motor
control of their extremities. Additionally a further 10.4% were also affected by
impairments in the spinal and torso region.
Functional loss of the use of extremities can be caused by upper spinal cord
trauma and degenerative diseases. Those afflicted are struggling, or are simply
unable, to perform every day tasks independently of others. One very prominent
area is the one of nutrition. Being self-sufficient in terms of being in control of
food and water intake is not only beneficial to ones health but also immensely
important for ones self-worth [30].
Assistive technologies are increasingly becoming a vital factor in the field of
assisted living; minimising the need for constant care and allowing people with
motor impairments to regain some independence [35]. Initial studies by Klein [24]
and Merkel & Kurcharski [34] indicated that assistive technology often meets
non-acceptance and non-use and propose that devices need to focus more on
the needs and preferences of the target group. Using a participatory approach
integrating future users in the developing progress is recommended to promote
a higher acceptance of the final product [50].
We conducted an ethnographic study in this work to shed light on how users
envision future systems supporting them with everyday drinking and eating
tasks. We interviewed 15 users with motor impairments, presented a robotic
aid as a potential assistive system, and conducted in-situ observations of their
drinking and eating behavior and used tools. We gained significant insight into
user opinions and derived recommendations regarding structural, social, and
collaborative concerns of future assistive systems like a robotic drinking aid
(cf., Figure 1. These recommendations will help designers and engineers in a
technology-focused domain to build systems that actually help people.
2 Related Work
Traditionally the focus in the field of developing assistive technologies has been
on functionality from an engineering point of view. Recent findings however
highlight the need to include future users and their perceived needs in the design
process [26]. In this section we first examine previous work done on concepts of
user participation and collaborative approaches. In a second step, we present
projects that already analysed the use of robotic devices to support people with
disabilities and how these aids are valued by their users.
2.1 User needs
In recent years, there has been growing interest in the concept of user partici-
pation in the design of new assistive technologies. Groundwork laid by Thielke

Recommendations for the Development of a Robotic Aid 3
et al. [48] and Merkel & Kucharski [34] expressed the need for this collaborated
approach to maximize user acceptance. They indicated various methods for in-
tegrating the user group as well as family, caregivers and assistants into the
innovation process. Focus groups, qualitative interviews, visits of the primary
users’ homes, and participant observation can provide significant insights into
the needs and wants of the user group. The recommendation for this participa-
tory approach that integrates the future users in the developing progress is also
noted by Frennert & ¨
Ostlund [17] and Eftring & Frennert [13], confirming the
findings by Klein and Merkel & Kurcharski.
During the development of a robotic therapy support system, Duckworth
et al. used three different methods to include the future users preferences into
their work [12]. Clinicians and patients were interviewed, given a questionnaire
concerning the design of a robotic therapy support system and had the oppor-
tunity to use the developed robot during counselling sessions. They came to
the conclusion that a participatory design provides essential information for the
development of assistive technology and increases the chance of a positive user
experience.
Using a similar approach, Mandy et al. conducted a qualitative study with
users of the Neater Eater to gain an in-depth understanding of their user ex-
perience [29]. They report that self-feeding devices increase the life quality of
people with disabilities significantly and support a more equal relationship be-
tween those who are in need of care and their carers. They stress the need of a
positive approach towards assistive technologies for a wide general acceptance.
2.2 Human-Robot Collaboration in the Field of Supporting People
with Disabilities
Robotic solutions can make a significant contribution to regaining independence
and improving care by supporting and relieving caregivers, thus improving the
quality of life of those in need of support [5].
A growing body of literature has examined the impact of assistive robotic
systems in supporting people with motor disabilities. Work done by Chen et
al. [9] for the Robots for Humanity project and Fattal et al. [14] looked into the
feasibility and acceptance of robotic systems as assistive technologies. A common
finding was that the robotic devices are often designed to assist with several
activities of daily living. These devices are usually large; consisting of a robotic
arm on a mobile module. They require a barrier-free environment and rooms
with sufficient space to fit into and be able to move around safely. In contrast,
Pascher et al. noted the potential of smaller, lightweight solutions designed for
individual tasks [36], indicating that a specialized aid would be more accessible
in terms of size and portability.
Research by Gallenberger et al. used camera and machine learning for an
autonomous robotic feeding system to detect types of food items present and to
plan the picking-up and transportation to the mouth of the user [18]. An alterna-
tive approach is presented by Canal et al. describing a learning-by-demonstration
framework to feed the user [8]. Both projects ensure the ability of the robotic

4 M. Pascher et al.
arm to fulfill its autonomous tasks without any fine-control of the user focusing
on the technical aspects of the development process of assistive technology.
A 2019 study by Beaudoin et al. focused on the long-term use of the robotic
arm JACO [4], a recent advance in assistive technologies. They researched im-
provements of everyday task capabilities, satisfaction with JACO, psychological
impact and the implications for users and their caregivers using a similar quanti-
tative approach as employed in this study. Beaudoin et al. reported that almost
all participants gained more autonomy in certain life aspects and experienced a
number of positive psycho-social impacts. One such success was the increased
capability of participants to drink independently of human support using JACO,
thus reducing the amount of care and attention needed and increasing well-being
and overall health by having a continuous access to beverages.
Interaction technologies such as gaze-based interaction and head movement
have been explored to operate, e. g. a PC [11,38,40] and a robot [22,41,45]. Alter-
natively, brain-computer interfaces were used to control a robotic arm [1]. How-
ever, today’s ubiquitous technology interaction scenarios are much more tightly
integrated in everyday activities and require different interaction interfaces [28].
3 Study
The goal of this work is to understand users’ requirements and demands of
assistive technology that supports them with drinking and eating. For this, we
conducted an ethnographic study consisting of an interview including a VR
presentation of a robotic support system and in-situ participatory observations
of their drinking and eating habits, with 15 participants afflicted by a varying
degree of motor impairments.
3.1 Participants
In preparation for the main study, we opted to evaluate our methods with a pilot
participant allowing us to adapt the study design before approaching the remain-
ing participants. Participants were chosen in collaboration with the Center for
Paraplegic Patients Hamburg, the Locked-in-Syndrom e.V. Berlin, and the State
Association of the German Society for Multiple Sclerosis Hessen e.V. We re-
cruited 15 participants with a permanent and significant degree of compromised
mobility of the extremities and the reliance on support for the consumption of
food and drinks. Table 1 presents the participants split by gender, age, and diag-
nosis. 4 female and 11 male participants took part in the main study; the mean
age was 42.07 years (SD=16.68) and all were categorized as severely disabled.
3.2 Procedure
Each session took place in the participants’ homes which allowed us to conduct
the interview, observation of drinking and eating habits as well as analysis of
commonly used aids in a natural setting. In most of the cases a caregiver or
assistant was present.

Recommendations for the Development of a Robotic Aid 5
ID Gender Age Diagnosis
Pilot female 60 Multiple sclerosis
P1 male 18 Spinal cord injury; incomplete at level C3
P2 male 46 Spinal cord injury; complete at level C4 & some
rudimentary mobility until level C5
P3 male 41 Spinal cord injury; incomplete at level C3 (right
body-side has some mobility until level C5)
P4 male 30 Spinal cord injury; incomplete at level C3
P5 female 62 Locked-in syndrom
P6 male 50 Spinal cord injury; incomplete at level C4
P7 male 38 Spinal cord injury; incomplete at level C3 & com-
plete at level C5
P8 male 30 Spinal cord injury; complete at level C3
P9 male 22 Spinal cord injury; complete from level C3 to C7
P10 male 48 Spinal cord injury; complete at level C4 & C5
P11 female 60 Multiple sclerosis
P12 male 50 Inclusion body myositis
P13 female 51 Locked-in syndrom
P14 male 34 Spinal cord injury; complete at level C5 & C6
P15 female 51 Arthrogryposis
Table 1: Overview of the pilot and main study participants
Interview Due to the nature of the physical impairments faced by the partici-
pants, obtaining their consent had to be adapted to their particular capabilities.
After reading or listening to a researcher reading the consent form, participants
signed the form by themselves or had their spoken agreement recorded. In other
cases, an authorized caregiver signed the form on behalf of the participant.
The interview part was structured in four sections; each focusing on a dif-
ferent aspect detailing their living situation, attitudes regarding drinking and
eating, level of assistance needed as well as wishes towards an ideal robotic aid.
In the first part we aimed to understand their current living situation by estab-
lishing how many hours they spend in their wheelchair, where they spend most
of the time, and where they eat and drink at home.
Next, we were interested in their value propositions and preferences regarding
drinking and eating. The participants were asked to describe a typical meal-
time routine, what they generally consume, and which preparations are needed.
Further, we wanted to know if drinking and eating is seen as a necessary task or
can also convey enjoyment. Participants were also asked if they consume food
and drinks if they are not at home (at work, in a restaurant).

6 M. Pascher et al.
The third step focused on the process of drinking and eating in an assistive
setting including the communication with their caregivers/assistants and any
improvised aids used.
In the final step of the session, we focused on the use of a proposed robotic
arm as a drinking and eating aid. To familiarize participants with the concept
they were shown images of different eating support systems and wheelchair-
extension-type robotic arms that are already on the market, e. g. iEat [3], Obi [10],
JACO [23], and iArm [2]. To simulate the situation of sitting in front of an ac-
tual robotic arm performing tasks in a close-contact environment we used Google
Cardboard [20] and a stereoscopic video of our in-lab robot setting. Conducting
the interviews in the participants’ homes made this lightweight solution nec-
essary. Figure 2 shows the robotic arm bringing a glass of water to the user’s
face (in this case simulated by the camera lens). To further the realism of the
situation, participants were able to experience the actual sounds of the robotic
aid by simultaneously listening to an audio recording.
Fig. 2: Stereoscopic video in a first-person perspective of sitting in front of a
robotic arm performing tasks in a close-contact environment
Following this experience we inquired about the participants’ perception of
the robotic aid including their likes and dislikes of the simulation and any changes
they would appreciate from an end-user perspective. We encouraged them to
express wishful thinking without worrying about current technological capabili-
ties. We were also interested in how the participants would like to interact and
collaborate with the robot. Special interest concerned the preferred location of
attachment (e.g. table, wheelchair or self-mobile) and which additional functions
should ideally be available.
Any additional thoughts, wishes, and suggestions of the participants were
recorded for use in future research.

Recommendations for the Development of a Robotic Aid 7
Participatory Observations This part of the study focused on observing
participants consume food and drinks (cf., Figure 3) with the assistance of their
caregiver. Observations of the relative location of the assistant, the methods
used, and the communication between both parties were recorded. Depending
on the specific type of impairment participants were either laying in their beds
or sitting in their wheelchairs. Filming these interactions allowed for easy access
during data analysis.
(a) Eating (front view) (b) Eating (back view)
(c) Drinking with a straw (d) Drinking with glass contact
Fig. 3: Observation of eating drinking habits together with their caregiver
Analysis of Commonly Used Aids During the qualitative interviews we
found that every study participant uses some kind of aid to facility food and drink
consumption (cf., Figure 4). In order to consume the necessary amount of fluids,
tea and water are provided in teapots (cf., Figure 4b) or in large dispensers (cf.,
Figure 4c). Large dispensers contain enough beverages for all day without the

8 M. Pascher et al.
need for re-filling by an assistant. We recorded images of these aids to increase
our understanding of the help people need and want when confronted with tasks
they cannot independently do anymore.
(a) Standard drinking straw
for a glass of water
(b) Drinking straw/hose for
warm and cold tea
(c) Filling glass with a dis-
penser
(d) Modification of ma-
chines (by adding parts)
(e) Self constructed bottle
opener
(f) Portable cutlery set
(g) Hydro flask (h) Mouth-stick rest (i) Lifter
Fig. 4: Analyzed aids in the participants’ homes which are currently used
3.3 Limitation of our study
The main target group of our study were people with quadriplegia caused by
spinal cord injury. The noticeable skewed ratio of more male than female partic-
ipants reflects statistic by the WHO [51] and German Federal Statistical Office

Recommendations for the Development of a Robotic Aid 9
(Destatis) [47] of a 2:1 male-to-female ratio for overall recorded spinal cord in-
juries worldwide [33]. Additionally, a higher number of women refused to take
part in our study as they felt uncomfortable with the study design (e. g. getting
filmed while eating and drinking).
Use of a Robotic Arm during the Study One participant (P15) has rudimentary
mobility functions in her lower right arm which allows her to use a joystick-
controlled robotic arm for nearly every activity of daily living including consum-
ing food and drinks, manipulating objects and basic hygiene. Figure 5 illustrates
how she handles the tasks with her robotic arm.
(a) Drinking (b) Eating
Fig. 5: Observation on the use of a robotic arm (JACO) for the consumption of
food and drinks
4 Results and Recommendations
Based on the analysis of interviews, observations, and images we devised a set of
user-centered design recommendations for a robotic drinking aid. Recommenda-
tions are split into three sections referring to structural, social and collaborative
concerns respectively. All materials were transcribed, coded, and categorized in-
dependently by two researchers. We focus on processes related to drinking and
eating, interactions between participants and caregivers, and additional topics of
interest. All participants were interviewed as experts in their own right, as they
can accurately describe and explain their situation, their abilities, limitations
and needs. Therefore, our analysis method for the interviews and questionnaires
followed the qualitative content analysis approach from Mayring [32]. Based on
a predetermined interview guide established by the research team, questions

10 M. Pascher et al.
belonging into different categories were discussed with study participants. De-
scriptive and normative statements concerning housing and living situation, in-
dividual wishes and needs regarding food and drink intake and attitude towards
robotic aids were analysed [7]. The analysis of videos and images was based on
the qualitative hermeneutical approach from social sciences by Reichertz & En-
glert [39] and the photo analysis by Pilarczyk & Mietzner [37]. In a first step, the
videos and photos were cataloged according to content (e. g. drinking aid) and
subject (e. g. drinking with a straw). Next, the videos were viewed, transcribed
and coded. In a last step the photos were viewed again and theme-oriented photo
series formed, e. g. photos showing self-made aids. Overall we conclude that pho-
tos provide additional information to the videos and interviews or can be used
to better describe findings but do not provide much value as standalone objects.
4.1 Structural Concerns
Any design process starts with a structural framework defining size, weight and
materials to be used. Whilst the choices might make sense from a purely technical
point of view the preferences of the end-user should still be considered. Nobody
benefits from the development of an assistive technologies that ends up too big
in size to be used in the home of the typical end-user. With the aim being the
widespread usage of the new device, taking wishes, where technically possible,
regarding size and design into considerations can only be beneficial to future
acceptance.
Dimensions of the Robotic Arm Although all but one of the participants
reported living in accessible housing, barriers including narrow hallways remain.
During the in-home session we found several of the participants housings to
be either too small for current robotic aids or lacking in space due to other
large assistive devices present. Care beds and tables, lifters and wheelchairs are
essential to support people with disabilities living in their own homes. Adding
another large-size device taking up space can be problematic and in some cases
impossible.
P10: “There is a second wheelchair somewhere, then maybe there is a bed-
side table somewhere, and there is a lifter somewhere and the shower chair
somewhere. (...) At some point, many run out of space.”
Recommendation 1
A robotic drinking aid should be primarily designed for saving space. The
arm has to have the ability to fold itself during waiting/suspended-mode.
And include the possibility of space-saving storage when not in use.
Physical Attachment Types of suitable attachment methods vary depending
on individual preferences, type of wheelchair used, and space availability in the

Recommendations for the Development of a Robotic Aid 11
participants’ homes. Frequent changes between user location (bedroom/living
room and bed/wheelchair) is a further factor to consider. Some participants use
a chin-controlled electronic wheelchair. This poses an additional challenge for
possible attachment methods and hinders both control and movement of the
robot during the drinking and eating process because the joystick is in this case
directly in front of the mouth.
P13: “It would be great to be able to fix the arm to the table with a small
screw clamp. (...) Adding it to the wheelchair might be good too. But in any
case it must be easy to dismantle.”
Recommendation 2
Different mounting options for the robotic arm have to be available to
allow attachment to different surfaces and care devices including mobility
aides, resting chairs or overbed / side tables. Special consideration has
to be given to the restrictions imposed by types of wheelchairs used.
4.2 Social Concerns
The advantage of an interdisciplinary approach as outlined in this study is the
combination between technical necessities and preferences of the end-user. Dur-
ing our analysis we found that the majority of respondents were much more
worried about ’social concerns’ than technical aspects.
Taking Design Seriously Stigmatization of people with disabilities is an on-
going problem highlighted in a number of studies and literature e. g. [31]. All
participants have reported that they worry about unwanted attention and further
stigmatisation by using too-conspicuous aids. Almost all asked for the robotic
arm to be unobtrusive and designed with a positive public image in mind.
P14: “I can imagine that design is relatively important, because it is likely
the crucial factor whether people accept it and whether they want to integrate
it into their environment, right?”
Recommendation 3
The design should range between something plain and unobtrusive to a
chic lifestyle product. The arm should be recognizable as a technical tool
and not mimic a human arm by using skin-colored coloration or skin-like
material.
The Care Situation and Social Aspects All participants relied on care from
in-home relatives for their daily needs. Additionally all but one also employed
professional personal care assistants. Due to the limited possibilities of outside
interactions the bond with family members and other regular caregivers was

12 M. Pascher et al.
observed as particularly strong and important for the mental well-being of the
participants.
Interviewer: “Would you describe the exchange between you and your assis-
tant during mealtime as formal or informal?” P4: “Very informal, just about
everyday life. Not just about mine. They tell me about themselves. And then
you just sit together and talk about everything that is going on. Daily events,
personal matters, politics, experiences, about everything really.”
Recommendation 4
Disruptions of conversations and social interactions by the drinking aid
have to be minimized. The robotic arm should be placed without ob-
structing the line of sight between user and assistant. Sound and noises
have to be kept to an absolute minimum to avoid distractions.
Safety Safe use must be guaranteed for primary and secondary users from the
onset. Teething problems must be avoided at all costs; therefore strict adher-
ence to safety protocols for direct physical proximity is vital. People within the
target group are already faced with numerous health concerns [6,46] and many
participants expressed worries about additional injury risks posed by the robot.
Worries surrounded their inability / decreased ability to move out of the way if
the robotic arm does not stop at a certain distance from their face. A frequently
suggested solution would be an adjustment to have the robot bringing a cup
with a straw close to the mouth - but not directly touching it; thus enabling the
user to cover the last few centimeters on their own accord.
Solutions include the aid of a straw to avoid the drinking cup being delivered
directly to the mouth. Allowing users the final approach increases their feeling
of autonomy, control and safety.
Interviewer: “What could prevent you from using the robotic arm?”
P2: “Teething troubles, something every device has at the beginning. If problems
with the programming come up and the whole weight of the robot would fall on
me.”
Recommendation 5
Apply the principle of safety first and design for scenarios of use avoiding
body-contact.
Privacy All participants require 24/7 assistance with results in very limited
privacy. They all stated that they have to drink a lot during the day for health
reasons. Particularly in the case of paraplegia, it is necessary to consume up to
three liters of fluid a day to support digestion and temperature regulation [21].
Being able to regulate fluid intake independently and not having to ask for
assistance every time they want to drink would allow users to spend several

Recommendations for the Development of a Robotic Aid 13
hours at a time without a caregiver. A frequently recorded hope concerns the
increase of time being alone gained by integrating the drinking aid into the users’
lives.
P4: “And you really sometimes want to be all alone. And even if I send
my assistants to go shopping and have an hour alone here or there, that’s not
comparable to really being alone. ”
Recommendation 6
Users have to be able to use the device, once set up, independently or with
minimal assistance. Once operational, assistants and caregivers should
not need to interact with the device at all. Potential components worn by
the user need to fit securely to prevent a constant need of re-adjustments
(cf., guidelines for wearability [19]).
Data Privacy and Security Only a small number of participants were con-
cerned about data protection. Some however expressed concerns about the type
of data collected, storage options, as well as access to it.
P6: “If there is a camera then I do not know where these images are going.
Especially if the robot is connected to the internet.”
Recommendation 7
Transparency about collected and stored data have to be maintained to
reduce uncertainty and skepticism about modern technology. Storage of
personal data, including camera images, should be avoided and frame-
works for voice commands should work offline as much as possible. If
data have to be stored, it has to be stored securely.
4.3 Collaborative Concerns
Effective assistive technologies only work if they can be used by the target group
without major effort. End-users know best what they are capable of and how
they feel most comfortable interacting with the robotic device. Therefore it is
important to consider the way they want to collaborate with their robotic aid.
Ease of Use The aim of using assistive technologies is increased independence of
the end-user; something that is only possible if the devices are easy and straight-
forward to use. Especially in the case of changing caregivers, it is exhausting for
users to repeatedly train others in the use of their robotic aid.
P1: “That means that I might not need a nurse anymore, but a technician.
Because I already struggle to instruct the carers; and that is just to trigger three
commands on my computer.”

14 M. Pascher et al.
Recommendation 8
Ease of use, preferably as ready-to-use design, should be the aim of the
assistive technology. Given the potential of frequently changing assistants,
intuitive design and an obvious command structure are required to ensure
a short - if at all necessary - familiarization periods. No prior knowledge
or training by secondary users can or should be expected and if anything
a short introduction guided by the primary user has to be sufficient. The
robotic arm should be - once adjusted to a mounting spot - ready to use
and easily used.
Interaction Design and Interaction Technology Participants in the study
indicated various desires regarding the interaction with the robotic aid. Due to
the frequent changes in position during the day, it is important that the robotic
arm is usable in a lying and in a sitting position either from a bed and from a
wheelchair.
The majority of the participants already use voice controlled components in
their homes, e. g. telephone, door opener, and lighting fixture. However, these
components generally cannot be compared with modern smart devices as they
do not connect to the internet. Only one participant used an smart speaker
for smart home solutions. Other participants refrained from using devices with
internet based voice control due to unreliable internet connections or - more
often - out of concerns towards data security.
P14: “So I think it would be great if it was using voice control. (...) I think
using a joystick or something similar is also very complex. But if I only have to
say: ‘Give me a glass of water’, and that would work, that would be great.”
Recommendation 9
Whilst voice control is preferred for control and interaction, speech im-
pairments must be taken into account with the extra requirements they
pose. Additionally, data security has been identified as a concern when
usage of internet based voice control is suggested. Offline solutions are
preferable to address these worries. Alternatively, eye-tracking devices
and data glasses can be viable options. Participants preferred the for-
mer two options compared to head gestures and headsets. Participants
preferred eye-tracking control via gaze-dwelling on either real world com-
ponents or virtual objects in combination with the data glass’s user in-
terface.
For users with residual hand and arm functions, a switch among semi-
autonomous mode and manual mode via direct joystick control is inter-
esting as it allows greater flexibility and adaptation to daily needs, due
to the fact that in a semi-autonomous mode scenarios have to been learnt
by the robot. The current mode has to communicated to the users and

Recommendations for the Development of a Robotic Aid 15
assistants, e. g. by a ring of two-colored LEDs around the robot’s flange
like a bracelet.
Robotic Arm as Combined Drinking and Eating Aid All Participants
are excited about the prospect of a functional drinking aid, allowing them to
independently regulate their fluid intake. In contrast, few participants can imag-
ine regularly using a robotic arm for food consumption. Those who can still eat
independently due to residual functionality in their upper extremities would like
to use and maintain this ability. A robotic arm as an eating aid is only interest-
ing for this group if food can be cut into small pieces with the help of the arm.
A cutting function would further increase their autonomy and enable participa-
tion in meal preparation; in their eyes another step towards social integration.
Participants who have their food served to them expressed satisfaction with the
assistance they receive from other people. They would like to continue in this
way because they value this social interaction and note that people can be more
flexible and spontaneous in responding to all eventualities. This includes emer-
gency situations such as choking or spillages, a worry of a number of participants
from our focus group voiced.
P5: “I would prefer [the aid] of my husband, because we do communicate a
little throughout lunch. I think when the robotic arm feeds you, there is just
silence.”
Recommendation 10
When prioritising the development of robotic assistance the first focus
should be on fluid intake. The scenario of eating with a robotic arm is
influenced by various complex aspects, such as social interactions, which
need further exploration.
Robotic Arm as a General Aid Participants frequently expressed a desire
for a robotic arm with a distinct grasping function beyond a mere drinking aid.
Desired functions include manipulating objects, such as picking something up,
taking something out of a cabinet, or being able to lift things. Fine motor tasks
such as turning the pages of a book or grasping easily breakable items were also
desired. In addition, particularly the younger participants would like to be able
to operate a game console. Furthermore, some of the participants would also like
to use a robotic arm for aspects of basic care, such as combing hair or brushing
teeth.
Participants who still eat independently also showed interest in the topic of
cooking. A robotic arm that can cut food, handle cooking utensils, and assist
with setting the table would increase autonomy and lead to more participation
in the entire process of eating. Participants expressed the wish to handle even
fragile objects like raw eggs or eat small but delicate snacks like crisps. Although

16 M. Pascher et al.
fears of possible stigmatization due to the use of the robotic arm exist, overall
the hope that a robotic arm with various functions could promote independent
and self-determined living whilst also giving relieve to caregivers was expressed.
P4: “Having such an arm fulfill different functions such as gripping, I think
that makes more sense because it would then be more versatile.”
Recommendation 11
Apart from functioning as a drinking and eating aid, a robotic arm
should be developed to fulfill other everyday tasks. Since participants
fear stigmatization over having too many tools, a robotic aid with vari-
ous functions would meet greater acceptance.
5 Discussion
Our ethnographic study provides recommendations for future research and de-
velopment as well as hypotheses that should be tested for further validation. In
the context of our target user group, implementing a solution based on our rec-
ommendations will still require adaptation to fit individuals with their specific
physical abilities, along with further research to verify that a designed assistive
system does indeed support the user. Recommendations regarding ”Taking De-
sign Seriously” and ”Privacy” concern interaction devices that might lead to
further stigmatization by drawing unwanted attention and require asking for as-
sistance for wearing or re-calibration - both aspects the target group wants to
avoid.
5.1 End-User Involvement in Assistive Technology
Assistive technologies are on the rise, with a number of different robotic aids
already on the market or in various stages of development [15,25]. Studies by
Scherer[43] and Verza[49] have shown that these devices, albeit useful in an
assistive setting, can have a high rate of non-acceptance and non-usage. There
is a growing body of literature indicating that this is due to the exclusion of the
end-user from the design process [16]. In recent years the field of collaborative
work between developers and end-users (or their advocates) has grown but is
still in its infancy as discussed by Lee et al.[27] and Simonsen [44]. Our work
represents such a collaborative approach investigating the needs and wants of the
end-user in regards to a robotic drinking and eating aid. In fact, participants
particularly valued the inclusion in the design process of a device developed
specifically for them.
5.2 Potential Autonomy
One important finding is that participants want the possibility of spending time
without their assistants. Specifically drinking as continuous hydration through-
out the day is vital for people with disabilities which results in a near-constant

Recommendations for the Development of a Robotic Aid 17
need of care when drinking independently is not possible [21]. Thus, the most
important capability of the robotic system should be to support the users with
drinking. On the other hand, however, the participants also noted that the num-
ber of assistive systems should be limited and, thus, the system should provide
multiple tasks. This trade off will be a core challenge for future developers.
5.3 Importance of Structural Concerns
Our results confirm findings by Fattal et al. [14] and others [34,48] as similar
structural concerns are also expressed by our participants. Thus, our results also
highlight the need for recommendations related to the physical characteristics
of the robotic arm and attachment site.
5.4 In-Home Methodology
In this study we recorded the preferences people with severe motor disabilities
have towards a robotic drinking and eating aid in terms of functionality and
design. We used in-home sessions to interview our participants and record their
everyday behaviours in a familiar setting. We opted for this particular approach
to increase authenticity of our observation in accordance with Sakowska [42].
Combining all findings from our conversations and observations allowed us to
gain significant insight in the actual living situation and challenges faced by the
target group. We believe the in-home methods used in this study to represent
a much more accurate picture than studies conducted in artificial laboratory or
workshop environments. One downside of this approach, however, is the limita-
tion to a small geographical area, potentially limiting the generalizability.
6 Conclusion
People with motor disabilities face a number of obstacles when confronted with
everyday tasks such as drinking and eating. Assistive technologies have the po-
tential to greatly improve the quality of life of the target group; however their
user acceptance has been challenged by previous work. In this paper we investi-
gated how drinking and eating aids are perceived by conducting interviews and
participatory observations. By analyzing the relationship with food and drink
intake as well as analyzing the wishes for future assistive technologies we were
able to better understand the needs and wants of the target group.
Our research has highlighted the importance of acknowledging structural, so-
cial, and collaborative concerns in respect to the design of a robotic arm, defining
a set of recommendations for the designs of robotic drinking aids. These recom-
mendations represent an important step in bridging the gap between technologi-
cal design and the preferences of the target group, thus increasing the likelihood
of acceptance of any further assistive technology.

18 M. Pascher et al.
Acknowledgement
We would like to thank all study participants and their assistants for their valu-
able opinions and time. The authors are also grateful for the support of the
Center for Paraplegic Patients Hamburg, the Locked-in-Syndrom e.V. Berlin,
and the State Association of the German Society for Multiple Sclerosis Hessen
e.V.. This research is supported by the German Federal Ministry of Education
and Research (BMBF, FKZ: 16SV7866K and 16SV7868).
References
1. Achanccaray, D., Chau, J.M., Pirca, J., Sepulveda, F., Hayashibe, M.: Assistive
robot arm controlled by a p300-based brain machine interface for daily activities.
In: 2019 9th International IEEE/EMBS Conference on Neural Engineering (NER).
pp. 1171–1174 (2019). https://doi.org/10.1109/NER.2019.8717042
2. Assistive Innovations B.V.: iARM — Robotic arm for humans, mountable on
powered wheelchair (02/05/2021), https://www.assistive-innovations.com/robotic-
arms/iarm
3. Assistive Innovations B.V.: iEAT Robot — Assistive feeding and eating robot for
people (02/05/2021), https://www.assistive-innovations.com/eatingdevices/ieat-
robot
4. Beaudoin, M., Lettre, J., Routhier, F., Archambault, P.S., Lemay, M.,
G´elinas, I.: Long-term use of the JACO robotic arm: a case series. Dis-
ability and Rehabilitation: Assistive Technology 14(3), 267–275 (Jan 2019).
https://doi.org/10.1080/17483107.2018.1428692
5. Bemelmans, R., Gelderblom, G.J., Jonker, P., de Witte, L.: Socially assistive
robots in elderly care: A systematic review into effects and effectiveness. Jour-
nal of the American Medical Directors Association 13(2), 114 – 120 (Feb 2012).
https://doi.org/10.1016/j.jamda.2010.10.002
6. Bickenbach, J.: International perspectives on spinal cord injury. World Health Or-
ganization, Geneva, Switzerland (2013)
7. Bogner, A.: Interviewing Experts - A practical Introduction (nach ”Interviews mit
Experten eine praxisorientierte Einf¨uhrung”). Springer VS, Wiesbaden (2014)
8. Canal, G., Aleny`a, G., Torras, C.: Personalization Framework for Adap-
tive Robotic Feeding Assistance. In: Social Robotics, pp. 22–31 (2016).
https://doi.org/10.1007/978-3-319-47437-3 3
9. Chen, T.L., Ciocarlie, M., Cousins, S., Grice, P.M., Hawkins, K., Hsiao, K., Kemp,
C.C., King, C.H., Lazewatsky, D.A., Leeper, A.E., Nguyen, H., Paepcke, A., Panto-
faru, C., Smart, W.D., Takayama, L.: Robots for humanity: using assistive robotics
to empower people with disabilities. IEEE Robotics Automation Magazine 20(1),
30–39 (Mar 2013). https://doi.org/10.1109/MRA.2012.2229950
10. DESiN LLC: Obi — The first dining robot of its kind (02/05/2021),
https://meetobi.com/
11. Duchowski, A.T.: Gaze-based interaction: A 30 year retrospective. Computers &
Graphics 73, 59–69 (Jun 2018). https://doi.org/10.1016/j.cag.2018.04.002
12. Duckworth, D., Henkel, Z., Wuisan, S., Cogley, B., Collins, C., Bethel,
C.: Therabot: The Initial Design of a Robotic Therapy Support Sys-
tem. In: Proceedings of the Tenth Annual ACM/IEEE International Con-
ference on Human-Robot Interaction Extended Abstracts. pp. 13–14 (2015).
https://doi.org/10.1145/2701973.2701993

Recommendations for the Development of a Robotic Aid 19
13. Eftring, H., Frennert, S.: Designing a social and assistive robot for se-
niors. Zeitschrift f¨ur Gerontologie und Geriatrie 49(4), 274–281 (May 2016).
https://doi.org/10.1007/s00391-016-1064-7
14. Fattal, C., Leynaert, V., Laffont, I., Baillet, A., Enjalbert, M., Leroux, C.: SAM,
an assistive robotic device dedicated to helping persons with quadriplegia: Us-
ability study. International Journal of Social Robotics 11(1), 89–103 (May 2018).
https://doi.org/10.1007/s12369-018-0482-7
15. Federici, S.: Assistive technology assessment handbook. CRC Press, Taylor & Fran-
cis Group, Boca Raton, FL (2018)
16. Federici, S., Corradi, F., Meloni, F., Borsci, S., Mele, M., De Sylva,
S., Scherer, M.: Successful assistive technology service delivery out-
comes from applying a person-centered systematic assessment pro-
cess: A case study. Life Span and Disability 18(1), 41–74 (Jul 2015),
http://www.lifespanjournal.it/client/abstract/ENG290 2.%20Federici.pdf
17. Frennert, S., ¨
Ostlund, B.: Review: Seven matters of concern of social robots and
older people. International Journal of Social Robotics 6(2), 299–310 (Jan 2014).
https://doi.org/10.1007/s12369-013-0225-8
18. Gallenberger, D., Bhattacharjee, T., Kim, Y., Srinivasa, S.S.: Transfer depends
on acquisition: Analyzing manipulation strategies for robotic feeding. In: 2019
14th ACM/IEEE International Conference on Human-Robot Interaction (HRI).
pp. 267–276 (2019). https://doi.org/10.1109/HRI.2019.8673309
19. Gemperle, F., Kasabach, C., Stivoric, J., Bauer, M., Martin, R.: De-
sign for wearability. In: Digest of Papers. Second International Sympo-
sium on Wearable Computers (Cat. No.98EX215). pp. 116–122 (1998).
https://doi.org/10.1109/ISWC.1998.729537
20. Google VR: Google Cardboard. https://arvr.google.com/cardboard/ (02/05/2021)
21. Haas, U.: Caring for People with Paraplegia: Problems, Needs, Resources and
Interventions (nach ”Pflege von Menschen mit Querschnittl¨ahmung Probleme,
Bed¨urfnisse, Ressourcen und Interventionen”). Huber, Bern (2012)
22. Jackowski, A., Gebhard, M., Thietje, R.: Head motion and head gesture-
based robot control: A usability study. IEEE Transactions on Neu-
ral Systems and Rehabilitation Engineering 26(1), 161–170 (Jan 2018).
https://doi.org/10.1109/TNSRE.2017.2765362
23. Kinova Inc: Kinova Jaco Assistive Robotic Arm (02/05/2021),
https://www.kinovarobotics.com/en/assistive-technologies/column-a1/kinova-
assistive-robotic-arm
24. Klein, B.: Aides, assistive Technologies, and Robotics: Maintaining Independence
and Quality of Life in old age (nach ”Hilfsmittel, Assistive Technologien und
Robotik: Selbstst¨andigkeit und Lebensqualit¨at im Alter erhalten”). Kohlhammer
Verlag (2020)
25. Kyrarini, M., Lygerakis, F., Rajavenkatanarayanan, A., Sevastopoulos, C.,
Nambiappan, H.R., Chaitanya, K.K., Babu, A.R., Mathew, J., Makedon,
F.: A survey of robots in healthcare. Technologies 9(1) (Jan 2021).
https://doi.org/10.3390/technologies9010008
26. Laitano, M.I.: Developing a participatory approach to accessible design. Interna-
tional Journal of Sociotechnology and Knowledge Development 9(4), 1–11 (Oct
2017). https://doi.org/10.4018/IJSKD.2017100101
27. Lee, H.R., ˇ
Sabanovi´c, S., Chang, W.L., Nagata, S., Piatt, J., Bennett, C., Hakken,
D.: Steps toward participatory design of social robots. In: HRI (ed.) HRI’17. pp.
244–253. IEEE, Piscataway, NJ (2017). https://doi.org/10.1145/2909824.3020237

20 M. Pascher et al.
28. Mahmud, S., Lin, X., Kim, J.: Interface for human machine interaction
for assistant devices: A review. In: 2020 10th Annual Computing and
Communication Workshop and Conference (CCWC). pp. 768–773 (2020).
https://doi.org/10.1109/CCWC47524.2020.9031244
29. Mandy, A., Sims, T., Stew, G., Onions, D.: Manual Feeding Device Ex-
periences of People With a Neurodisability. American Journal of Oc-
cupational Therapy 72(3), 7203345010p1–7203345010p5 (May 2018).
https://doi.org/10.5014/ajot.2018.025353
30. Martinsen, B., Harder, I., Biering-Sorensen, F.: The meaning of assisted feeding
for people living with spinal cord injury: a phenomenological study. Journal of
Advanced Nursing 62(5), 533–540 (May 2008). https://doi.org/10.1111/j.1365-
2648.2008.04637.x
31. Matera, C., Nerini, A., Di Gesto, C., Policardo, G.R., Maratia, F., Dalla Verde,
S., Sica, I., Paradisi, M., Ferraresi, L., Pontvik, D.K., Lamuraglia, M., March-
ese, F., Sbrillo, M., Brown, R.: Put yourself in my wheelchair: Perspective-
taking can reduce prejudice toward people with disabilities and other stigma-
tized groups. Journal of Applied Social Psychology 51(3), 273–285 (Mar 2021).
https://doi.org/10.1111/jasp.12734
32. Mayring, P.: Qualitative Content Analysis: Basics and Techniques (nach ”Quali-
tative Inhaltsanalyse: Grundlagen und Techniken”). Beltz, Weinheim (2015)
33. McColl, M.A., Charlifue, S., Glass, C., Lawson, N., Savic, G.: Aging, gender, and
spinal cord injury. Archives of Physical Medicine and Rehabilitation 85(3), 363–
367 (Mar 2004). https://doi.org/10.1016/j.apmr.2003.06.022
34. Merkel, S., Kucharski, A.: Participatory Design in Gerontechnology: A Sys-
tematic Literature Review. The Gerontologist 59(1), e16–e25 (May 2018).
https://doi.org/10.1093/geront/gny034
35. Park, D., Hoshi, Y., Mahajan, H.P., Kim, H.K., Erickson, Z., Rogers, W.A., Kemp,
C.C.: Active robot-assisted feeding with a general-purpose mobile manipulator:
Design, evaluation, and lessons learned. Robotics and Autonomous Systems 124,
103344 (Feb 2020). https://doi.org/10.1016/j.robot.2019.103344
36. Pascher, M., Schneegass, S., Gerken, J.: SwipeBuddy: A Teleoperated Tablet and
Ebook-Reader Holder for a Hands-Free Interaction. In: Lamas, D., Loizides, F.,
Nacke, L., Petrie, H., Winckler, M., Zaphiris, P. (eds.) Human-Computer Interac-
tion – INTERACT 2019. pp. 568–571. Springer International Publishing, Cham
(2019). https://doi.org/10.1007/978-3-030-29390-1 39
37. Pilarczyk, U., Mietzner, U.: Picture science methods in research on ed-
ucation and social science (nach ”Bildwissenschaftliche Methoden in der
erziehungs- und sozialwissenschaftlichen Forschung”). Zeitschrift f¨ur qualita-
tive Bildungs-, Beratungs- und Sozialforschung 1(2), 343–364 (Jan 2000),
https://www.ssoar.info/ssoar/handle/document/28057
38. Plaumann, K., Ehlers, J., Geiselhart, F., Yuras, G., Huckauf, A., Rukzio, E.: Bet-
ter than you think: Head gestures for mid air input. In: Human-Computer Interac-
tion — INTERACT 2015, pp. 526–533. Springer International Publishing (2015).
https://doi.org/10.1007/978-3-319-22698-9 36
39. Reichertz, J.: Introduction to Qualitative Video Analysis — A hermeneutic-
sociological Analysis Tool (nach ”Einf¨uhrung in die qualitative Videoanalye —
Eine hermeneutisch-wissenssoziologische Fallanalyse”). VS Verlag f¨ur Sozialwis-
senschaften, Wiesbaden (2011)
40. Roig-Maim´o, M.F., MacKenzie, I.S., Manresa-Yee, C., Varona, J.: Head-tracking
interfaces on mobile devices: Evaluation using fitts’ law and a new multi-directional

Recommendations for the Development of a Robotic Aid 21
corner task for small displays. International Journal of Human-Computer Studies
112, 1–15 (Apr 2018). https://doi.org/10.1016/j.ijhcs.2017.12.003
41. Rudigkeit, N., Gebhard, M.: Amicus—a head motion-based interface for control of
an assistive robot. Sensors 19(12) (Jun 2019). https://doi.org/10.3390/s19122836
42. Sa lkowska, M.: Carrying out research among persons with disabilities and their
relatives — selected ethical issues. Zoon Politikon (8), 200–217 (Dec 2018).
https://doi.org/10.4467/2543408XZOP.18.010.10066
43. Scherer, M.J.: Living in the state of stuck: How assistive technology impacts the
lives of people with disabilities. Brookline Books (2005)
44. Simonsen, J.: Routledge international handbook of participatory design. Rout-
ledge, New York (2013)
45. Stalljann, S., W¨ohle, L., Sch¨afer, J., Gebhard, M.: Performance analysis of a head
and eye motion-based control interface for assistive robots. Sensors 20(24) (Dec
2020). https://doi.org/10.3390/s20247162
46. Statistisches Bundesamt (Destatis): Statistical Yearbook 2019 — Chapter
4 Health (nach ”Statistisches Jahrbuch 2019 — Kapitel 4 Gesundheit”).
https://www.destatis.de/DE/Themen/Querschnitt/Jahrbuch/jb-gesundheit.pdf
(2019)
47. Statistisches Bundesamt (Destatis): Disability Facts and Figures –
Brief Report 2019 (nach ”Statistik der schwerbehinderten Menschen –
Kurzbericht 2019”). https://www.destatis.de/DE/Themen/Gesellschaft-
Umwelt/Gesundheit/Behinderte-Menschen/Publikationen/Downloads-
Behinderte-Menschen/sozial-schwerbehinderte-kb-5227101199004.html (2020)
48. Thielke, S., Harniss, M., Thompson, H., Patel, S., Demiris, G., Johnson, K.:
Maslow’s Hierarchy of Human Needs and the Adoption of Health-Related Tech-
nologies for Older Adults. Ageing International 37(4), 470–488 (Dec 2012).
https://doi.org/10.1007/s12126-011-9121-4
49. Verza, R., Carvalho, M.L.L., Battaglia, M.A., Uccelli, M.M.: An interdis-
ciplinary approach to evaluating the need for assistive technology reduces
equipment abandonment. Multiple Sclerosis Journal 12(1), 88–93 (Feb 2006).
https://doi.org/10.1191/1352458506ms1233oa
50. Vines, J., Clarke, R., Wright, P., McCarthy, J., Olivier, P.: Configuring partic-
ipation: On how we involve people in design. In: Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems. pp. 429–438 (2013).
https://doi.org/10.1145/2470654.2470716
51. World Health Organization: Spinal cord injury (02/05/2021),
https://www.who.int/news-room/fact-sheets/detail/spinal-cord-injury