SAPA Technology: An AAL Architecture for Telemonitoring
Hubert Ngankam
, Maxime Lussier
2 b
, Aline Aboujaoud´e
2 c
, C´edric Demongivert
3 d
H´el`ene Pigot
, S´ebastien Gaboury
3 f
, Kevin Bouchard
3 g
, M´elanie Couture
4 h
, Nathalie Bier
2 i
and Sylvain Giroux
Laboratoire DOMUS, D´epartement d’Informatique, Universit´e de Sherbrooke, Sherbrooke, Canada
Centre de Recherche de l’Institut Universitaire de G´eriatrie de Montr´eal - CIUSSS-CSMTL,
Universit´e de Montr´eal, Montr´eal, Canada
LIARA Lab, Universit´e du Qu´ebec `a Chicoutimi, Chicoutimi, Canada
ES, CIUSSS West-Central Montreal, Cˆote-Saint-Luc, Canada
{nathlie.bier, aline.aboujaoude},,
{cedric.demongivert1, Sebastien
Gaboury, Kevin Bouchard}
Ambient Assisted Living, Event-driven Architecture, Event Streaming, Apache Kafka, Spark, IoT, ADL,
Lambda Architecture.
Ambient Assisted Living (AAL) aims to allow frail older adults to stay safe at home, partly through remote
monitoring which offers clinicians a means to prevent and manage risks. AAL needs an architecture to support
the large set of data emanating from multiple sensors dispatched in several smart homes. These data must be
processed in real-time to take the appropriate decisions in time. In this article, we propose an Event-Driven
Architecture according to the publish-subscribe pattern. The proposed architecture is the core of our system,
named SAPA Technology. It is composed of three layers: data gathering, data ingestion, and data processing.
To ingest the data stream, we choose Apache Kafka, an open-source broker, and Apache Spark, a streaming
system to process the ingested data. The SAPA Technology architecture respects scalability, homogeneity, and
modularity. It supports at least thirty-eight smart homes.
The aging of population in Canada and elsewhere in
the world changes the way social and health services
must be delivered (Organization, 2015). Only 10% of
Canadian over 65 years-old live in a retirement home,
or in a nursing home if their health status necessi-
tates more medical services (Roy et al., 2018). But
the number increases when incapacities and manage-
ment of multiple chronic disease necessitate moving
into a nursing home, or relying on increase home-care
services. Even in the context of disability, older adults
express the will to stay at home for as long as possible.
However, some diseases can lead to severe cognitive,
physical, and sensory deficits such as neurodegener-
ative diseases, strokes, arthritis, or macular degener-
ation. These conditions can have significant impacts
on functional independence, especially in Activities
of Daily Living (ADL). They may cause multiple risk
situations whether immediate, such as falls, or on the
long term, such as social isolation, malnutrition, or
poor hygiene (Boulos et al., 2017; Robinson, 2018).
Research into Ambient Assisted Living (AAL) fo-
cuses on automated methods or processes to support
safe performance in ADL, particularly for older adults
with multiple health conditions while being supported
by family, friends and medical staff (Rashidi and Mi-
hailidis, 2013). ALL involvea sensitive, adaptive, and
responsive software system to changes in the condi-
tions of the persons living in the smart homes.
One of the AAL system aims is to support the clin-
ical decision-making of clinicians delivering home
Ngankam, H., Lussier, M., Aboujaoudé, A., Demongivert, C., Pigot, H., Gaboury, S., Bouchard, K., Couture, M., Bier, N. and Giroux, S.
SAPA Technology: An AAL Architecture for Telemonitoring.
DOI: 10.5220/0010973400003123
In Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 5: HEALTHINF, pages 892-898
ISBN: 978-989-758-552-4; ISSN: 2184-4305
2022 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
care services (Lussier et al., 2020c; Lussier et al.,
2020a). To do this, the AAL system monitors ADL
performance and changes in habits and produce re-
ports to inform clinicians. Clinical decisions re-
lated to the services that should be delivered to the
older adults are then based on objective informa-
tion (Lussier et al., 2020c). Offering periodic reports
thus requires mechanisms for collecting and analyz-
ing real-time data coming from smart homes.
Usually, data collection is done using sensors,
connected objects and Internet of Things (IoT) de-
vices capable of responding to an environmentalstim-
ulus. Thanks to this equipment, monitoring (Ka-
nis et al., 2013; Zaric et al., 2015), context aware-
ness (Wood et al., 2008; Schilit et al., 1994) and activ-
ity recognition (Charlon et al., 2013;
et al., 2013) can offer real-time services. The con-
tinuous flow of data emanating from the smart home
requires computer methods to process them within a
reasonable time frame. Above all, rapid data process-
ing provides to the clinicians means to take informed
The great diversity of sensors and smart devices
available implies to set up an infrastructure that man-
ages heterogeneous data. Each device may contain
several sensors with one or more parameters, which
results in the creation of heterogeneous data. The
AAL architecture must then manage the real-time
streaming data emitted by these sensors as well as ap-
plications for data ingestion and processing to provide
appropriate services to the older adults and the clini-
cal staff.
To support the independence of older adults,
the health care system in Qu´ebec, Canada, has
implemented a program called SAPA (Soutien `a
l’Autonomie des Personnes
Ag´ees; Support for the
Autonomy of the Older Adults). SAPA Technology
is a project conducted by our team in close collabo-
ration with SAPA programs in Qu´ebec, that aims to
promote home care services for isolated older adults
with cognitive deficits through the use of remote mon-
itoring of ADL. It aims to assess use cases and usabil-
ity of the monitoring system. Ultimately, the project
aims to support home care services in delivering the
right services, for the right person, at the right time.
This article mainly deals with the technological aspect
of the SAPA Technology project. It presents, through
software and hardware architecture, how remote mon-
itoring is constructed, to facilitate home care through
the analysis and treatment of ADL.
This article presents a scalable architecture opti-
mized for real-time and batch data processing. Based
on the Lambda architecture approach (Kiran et al.,
2015), the proposed architecture combines the capa-
bilities of collection, ingestion and batch and flow
processing to recognize ADL, such as hygiene, meals
preparation, rest, outings and inactivity. Lambda ar-
chitecture is a data-processing architecture designed
to handle massive quantities of data by taking advan-
tage of both batch and stream-processing methods.
Driven by events, the proposed model subscribes to
IoT equipment events to offer to clinicians an ADL
analysis interface.
The remainder of the article is structured as fol-
lows: the clinical needs for telemonitoring are pre-
sented in Section 2; the SAPA Technology architec-
ture model based on Lambda architectures is pre-
sented in section 3; the tools used to implement the ar-
chitecture are presented in section 4; the whole SAPA
Technology system and some clinical elements are
discussed in Section 5; and, finally, our conclusions
are given in section 6.
AAL is a promising alternative to promote aging at
home and prevent risks. Remote monitoring offers
means to the clinicians to be warned when a haz-
ardous event occurs, and to determine social and med-
ical services needed when the health status of the
older adults evolves (Lussier et al., 2020b). Thanks
to the sensors scattered around the person’s home,
precise information about their everyday life routine
is collected. They can be combined with informa-
tion reported by the older adults themselves, and/or
their family, about how ADL are performed. Among
those, the most important ADL are meals preparation,
hygiene, sleep and outings (Rashidi and Mihailidis,
2013; Pollak and Perlick, 1991; Roy et al., 2016). The
global level of indoors activity complements these
ADL, since inactivity informs on the potential need
of interventions regarding a health condition or moti-
The expectations of the SAPA stakeholders taking
part in the study is to be able to better understand the
everyday life routine of the older adults with cogni-
tive deficits and to obtain specific information about
some of the activities they are engaging in. To do so,
the SAPA Technology provides reports on the activ-
ities occurring at the older adults home to the clin-
ician who may then decide to modify, or not, their
intervention plan. Thereafter, the clinician can follow
the intervention plans’ implementation through regu-
lar reports offered by SAPA Technology (Figure 1).
This remote monitoring of ADL is based on an AAL
environment deployed at the older adults’ home and
SAPA Technology: An AAL Architecture for Telemonitoring
gathering continuous information about activities oc-
curring at home (i.e, meal preparation, hygiene, sleep,
outings and periods of inactivity).
Figure 1: SAPA Telemonitoring process.
By their nature, AAL systems are real-time systems
that monitor the performance of ADL with the aim
to assist when needed. An AAL system therefore in-
volves the modeling of a real-time architecture.
This real-time prerequisite implies that the ambi-
ent sensors operate most often in event mode, namely
in publish-subscribe patterns. According to this pat-
tern, every detection of an action in the smart home
triggers a sensor event (Kenfack Ngankam et al.,
2020). An architecture that follows the publish-
subscribe pattern is named Event-DrivenArchitecture
(EDA) (Chandy, 2006). This approach contrasts with
the SOA architecture approach that gathers data for
responding to requests (Hanson, 2005).
The subscription model is better suited to push
events to consumers rather than to offer passive data
read mechanisms. With each significant change in the
home, EDA promotes production, propagationand re-
action to events. Thus, an event architecture favours
responsiveness of the AAL system, as systems based
on events are designed purposely for unpredictable
and asynchronous environments (Hanson, 2005). Re-
leased from the data storage constraint imposed by re-
quests, an event architecture greatly contributes to the
scalability and adaptability of AAL systems. The re-
mainder of the section first presents the criteria taken
into account for modeling the EDA and, thereafter,
the three components constituting the SAPA Technol-
ogy architecture: data gathering, data ingestion and
data processing.
3.1 EDA Requirements
The SAPA Technology architecture must deliver in-
formation to the clinicians, and assistance to the older
adults, on a real-time basis, but it must also guarantee
reliability and security for all the analyses and pro-
cesses realized. Therefore, the EDA must satisfy three
types of specifications:
1. Respond to the needs of final users
(a) Adaptable to the different needs of the older
(b) Non-intrusive to facilitate the acceptability.
(c) Easy to use and offer real-time data stream pro-
cessing capability.
2. Be devices independent
(a) Homogeneous to support different manufactur-
ers and protocols in the same network.
(b) Generic enough to be independent of commu-
nication technology.
(c) Suitable for different smart environments.
3. Be extensible to integrate new services
(a) Modular, stand-alone and based on reusable
(b) Scalable to allow new sensors and components
to be added or removed.
Lambda architecture is a data processing archi-
tecture, which takes the advantages of both batch
processing and stream-processing to handle a large
amount of data effectively (Kiran et al., 2015). The
SAPA Technology architecture is a model based on
Lambda architecture and consists of three layers. The
data gathering layer is fully distributed and dissemi-
nated in the environment. Data ingestion occurs both
at the edge and in the cloud to enable better interoper-
ability. In the third layer, processing and application
are performed in the cloud to benefit from the perfor-
mance of compute tools.
3.2 Data Gathering
Sensors and IoT devices gather smart homes events
to identify ADL. The interfacing of specific sen-
sors/devices with the system is achieved by a local,
distributed and autonomous data collection compo-
nent. This component acts as an intermediate layer
aimed at clearly separating IoT devices from the rest
of the system. This separation makes it possible to
increase the modularity and to increase the weak cou-
pling between the components of the architecture, as
shown in Figure 2. To facilitate context awareness,
and promote the efficiency of exchanges, the privi-
leged communications are of the machine-to-machine
(M2M) type. The data collected by the sensors is
transmitted to a local processing unit and to a data
stream ingestion unit in real-time.
Smart CommuniCare 2022 - Special Session on Smart Living Environments to Support Aging-in-Place in Vulnerable Older Adults
Figure 2: SAPA Lambda Architecture.
3.3 Data Ingestion
We propose a heterogeneous data ingestion model, as
shown in Figure 2. The model can receive or extract
heterogeneous data from multiple sources and save it
in a unified format. Built to do data joins, aggrega-
tions, filters and transforms, the data ingestion model
is based on distributed event streaming. Event stream-
ing generates a continuous flow and interpretation of
data to ensure that the right information is in the right
place at the right moment. Four strategies based on
high performance data pipelines are used to model the
unit of ingestion. They are high availability, scalabil-
ity, high throughput, and permanent storage.
In the SAPA Technology architecture, data is in-
gested from multiple different data sources at the
same time. Each source submits new tuples on a
stream, which are then received by the data ingestion
mechanism. This data ingestion primarily serves as
a mail queue. It routes the tuples to the correct des-
tination while continuously triggering the appropriate
Extract, Transform, and Load (ETL) process as new
data arrives. The publish and subscribe process that
analyses and transforms data ensures interoperability
and reusability. The ingestion of each event does not
require a response to the entity that delivered the data.
It is a one-way data pipeline.
3.4 Data Processing
As a Lambda architecture, the SAPA Technology ar-
chitecture was designed to ingest and process data in
real-time (Figure 2). The data processing unit is a
software component working either asynchronously
or synchronously. It detects independent incoming
events of different types and identifies a high-level
event (e.g., ADL, by correlating these sensor events
with the others). In this sense, high-level events
can be defined as the output generated after process-
ing many small IoT independent input data streams.
High-level events can be: sleep, rest, meal prepara-
tion, hygiene, outings and periods of inactivity (De-
mongivert et al., 2021; Demongivert et al., 2020).
Data coming from data ingestion is sent both to
the batch processing layer and the speed layer for pro-
cessing. Batch processing uses a batch query to gen-
erate analyses and to identify high-level events. The
main function of the batch layer is to use a histori-
cal archive to keep all the data collected. The speed
layer is similar to the batch layer. It computes also
similar analyses, except these analyses are realized in
real-time and concern only the most recent data. To
achieve this, it uses the queuing and streaming mech-
anisms. The flexibility of such an architecture makes
it easy to adapt to the scalability of the older adult’s
The implementation of the SAPA Technology archi-
tecture was primarily driven by the real-time nature
of IoT systems. It integrates hardware and software
components to support telemonitoring services. All
the sensors of the SAPA Technology architecture use
the Z-Wave and ZigBee communication protocols to
transmit data.
Mainly due to the event-driven nature of IoT sen-
sors, the implementation favored transmission in pub-
lish and subscribe mode. The publication and sub-
scription mechanism involves three types of compo-
nents: a client that sends messages, called a pub-
lisher, a second client that receives messages, called
SAPA Technology: An AAL Architecture for Telemonitoring
a subscriber, and a broker, responsible for manag-
ing the communication. MQTT (Message Queuing
Telemetry Transport) (Hunkeler et al., 2008) is a
lightweight network publish and subscribe protocol
based on the Transmission Control Protocol (TCP).
MQTT is open, simple, and easy to implement on de-
vices with limited resources, in environments such as
M2M communication and IoT. The MQTT broker is
the main component of the data collection unit. A hub
module is used as edge computing for local process-
ing and ambient decision making.
A streaming system needs a messaging infrastruc-
ture at the start of its pipeline to feed the system with
data. In the context of AAL systems, messages typi-
cally indicate the state of an IoT device at any given
Apache Kafka (Kreps et al., 2011) is an open-
source message broker and highly scalable publish
and subscribe messaging system capable of handling
thousands of clients and hundreds of megabytes of
read and write per second. Kafka emphasizes high
speed messaging, scalability and sustainability. Kafka
supports both batch consumers which may be offline
and online consumers which require lowlatency. Well
suited to equipment, subject to strong resource con-
straints, Kafka can manage long message delays to
provide periodic system ingestion. It allows con-
sumers to replay messages as needed, it is capable
of queuing new tuples or pushing new ETLs to the
streaming processing unit. This functionality is im-
portant for our architecture that is shared by multiple
services. Kafka was chosen for these two reasons.
The automatic failover is realized by failure de-
tection and active node election mechanism imple-
mented on a distributed configuration service with
Apache Zookeeper. Currently, in the SAPA Tech-
nology implementation, Kafka serves exclusively as
a message queue for individual tuples, each of which
is routed to the appropriate data flow graph in the
streaming processing engine.
The Lambda architecture was proposed by Marz
and Warren (Warren and Marz, 2015) to provide a
scalable and fault-tolerant architecture for processing
real-time and historical data in an integrated manner.
It is the pillar of the SAPA Technology architecture to
analyze large amounts of data in an efficient, fast, and
fault-tolerant manner.
The SAPA Technology architecture uses
Spark (Zaharia et al., 2010), a modern stream-
ing system offering highly scalable processing with
low latency. Spark provides data-driven processing,
batch processing, and streaming primitives, all of
which seem like a natural fit for SAPA Technology.
Spark is specially designed for latency sensi-
tive applications that involve large volumes of data
streams with time stamps, such as trading systems,
fraud detection and surveillance applications. Spark
provides the ability to perform in-memory calcula-
tions using resilient distributed data sets (RDDs), al-
lowing it to provide faster compute times for iterative
applications. Spark streaming processes data streams
in micro-batches, where each batch contains a collec-
tion of events that occurred during the batch period
(regardless of when the data was created).
The SAPA Technology is currently installed in
thirty-eight smart homes. In each of these smart
homes, an older adult is followed by a clinician who
regularly carries out home care services follow-ups
to adapt the range of services needed. The SAPA
Technology has been validated and deployed in ac-
cordance with the Ministry of Health and Social Ser-
vices’ rules and regulations regarding data security
and protection. Each house contains an average of
thirty-eight connected devices. Typically, devices are
composed of one to six built-in sensors. The SAPA
Technology architecture currently supports more than
three thousand sensors and processes, eight hundred
raw events per minute for the detection of more than
twelve high-level events (rest, meal preparation, out-
ings, etc.) per minute (Figure 3).
The passage from a SOA architecture to a Lambda
architecture allows to process a large amount of data
coming from multiple smart homes. Therefore, the
analysis is made easily on real-time or on batch mode
to provide in depth analysis. The scalability con-
cerns the number of devices installed in the homes,
the number of smart homes and also the number of
clinicians who could observe the data. The SAPA
Technology architecture is also less prone to faults
as the archive offers means to recover false analy-
sis. The three layers architecture offers interoperabil-
ity as communications between the layers have been
judiciously chosen to ensure that various components
may communicate together through an intermediate
The SAPA Technology architecture offers modu-
larity thanks to the Kafka pipeline that enables ser-
vices to share data and the analysis they have made.
Clinicians are able to visualize ve ADL on a secure
web platform. Regular reports are available to give
indicators on ADL to support their clinical decision-
making regarding services needed. Thanks to the
modularity, we also plan to add services to the older
adults in order to assist them and to provide a safe en-
Smart CommuniCare 2022 - Special Session on Smart Living Environments to Support Aging-in-Place in Vulnerable Older Adults
Figure 3: SAPA Lambda Architecture Performance.
vironment without changing the internal code struc-
ture or the underlying technology.
The deployment of the SAPA Technology in a
smart home necessitates installing the physical de-
vices at home and to linking them to the SAPA archi-
tecture. In our architecture, services for the installer
facilitate and shorten the time to deploy the sensors at
This article has presented the architecture that pro-
vides remote monitoring in an AAL system for older
adults. This research on computer architecture is part
of a larger research program conducted by a transdis-
ciplinary team composed of computer scientists, oc-
cupational therapists, psychologists and designers. It
is the cornerstone of multiple services for older adults
and clinicians as well as for the research program.
Without a reliable architecture that allows to
gather accurate information and to provide appropri-
ate reasoning in real-time, clinicians and researchers
may not be able to conduct other research programs
and to develop services. This architecture is the result
of the transdisciplinary team as all the needs identi-
fied by each discipline have contributed to elaborate
the specifications of the architecture to attain a com-
mon goal.
In a near future, we plan to add services to the
older adults, such as a calendar to remember appoint-
ments and planned activities. Thanks to SAPA Tech-
nology it will be easy to link the additional services
to the existing ones in order for the applications to
share the same data. To identify the ADL occurring
in the smart home, SAPA Technologies infer the ADL
recognition from rules and cross-referencing of event
data. It is planned to use data mining approaches,
which are available in Spark, to better follow the evo-
lution of the older adults autonomy.
Special thanks to the DOMUS laboratory develop-
ment team who spent several weeks developing and
testing this architecture, in particular Paul Guer-
lin - Yannick Drolet - Mauricio Chiazzaro - Math-
ieu Gagnon. The research is funded by the AGE-
WELL network and the Collaborative Health Re-
search Projects of Canada. Nathalie Bier is supported
by a salary award from the Fonds de la recherche du
Qu´ebec - Sant´e.
Alvarez-Garc´ıa, J. A., Barsocchi, P., Chessa, S., and Salvi,
D. (2013). Evaluation of localization and activity
recognition systems for ambient assisted living: The
experience of the 2012 EvAAL competition. Jour-
nal of Ambient Intelligence and Smart Environments,
Boulos, C., Salameh, P., and Barberger-Gateau, P. (2017).
Social isolation and risk for malnutrition among older
people. Geriatrics & Gerontology International,
Chandy, K. M. (2006). Event-driven applications: Costs,
benefits and design approaches. Gartner Application
Integration and Web Services Summit, 2006.
Charlon, Y., Bourennane, W., Bettahar, F., and Campo, E.
(2013). Activity monitoring system for elderly in a
context of smart home. IRBM, 34(1):60–63.
Demongivert, C., Bouchard, K., Gaboury, S., Lussier, M.,
Kenfack-Ngankam, H., Couture, M., Bier, N., and
Giroux, S. (2020). Fuzzy Features for Quality Estima-
tion of Activity Instances in a Dataset. In Proceedings
of the 6th EAI International Conference on Smart Ob-
jects and Technologies for Social Good, pages 54–59,
New York, NY, USA. ACM.
Demongivert, C., Bouchard, K., Gaboury, S., Lussier, M.,
Kenfack-Ngankam, H., Couture, M., Bier, N., and
Giroux, S. (2021). Handling of Labeling Uncer-
tainty in Smart Homes using Generalizable Fuzzy
Features. In Proceedings of the Conference on Infor-
mation Technology for Social Good, pages 248–253,
New York, NY, USA. ACM.
Hanson, J. (2005). Event-driven services in SOA. Java-
world, January, 31.
SAPA Technology: An AAL Architecture for Telemonitoring
Hunkeler, U., Truong, H. L., and Stanford-Clark, A. (2008).
MQTT-S—A publish/subscribe protocol for Wire-
less Sensor Networks. In 2008 3rd International
Conference on Communication Systems Software and
Middleware and Workshops (COMSWARE’08), pages
Kanis, M., Robben, S., Hagen, J., Bimmerman, A., Wage-
laar, N., and Kr¨ose, B. (2013). Sensor Monitoring in
the Home : Giving Voice to Elderly People. Perva-
sive Computing Technologies for Healthcare (Perva-
siveHealth), 2013 7th International Conference on.
Kenfack Ngankam, H., Pigot, H., Lorrain, D., Viens,
I., and Giroux, S. (2020). Context awareness ar-
chitecture for ambient-assisted living applications:
Case study of nighttime wandering. Journal of Re-
habilitation and Assistive Technologies Engineering,
Kiran, M., Murphy, P., Monga, I., Dugan, J., and Baveja,
S. S. (2015). Lambda architecture for cost-effective
batch and speed big data processing. In 2015 IEEE In-
ternational Conference on Big Data (Big Data), pages
Kreps, J., Narkhede, N., Rao, J., and others (2011). Kafka:
A distributed messaging system for log processing. In
Proceedings of the NetDB, volume 11, pages 1–7.
Lussier, M., Aboujaoud´e, A., Couture, M., Moreau, M.,
Lalibert´e, C., Giroux, S., Pigot, H., Gaboury, S.,
Bouchard, K., Belchior, P., Bottari, C., Par´e, G., Con-
sel, C., and Bier, N. (2020a). Using Ambient Assisted
Living to Monitor Older Adults With Alzheimer Dis-
ease: Single-Case Study to Validate the Monitoring
Report. JMIR Medical Informatics, 8(11):e20215.
Lussier, M., Aboujaoud´e, A., Couture, M., Moreau, M.,
Lalibert´e, C., Giroux, S., Pigot, H., Gaboury, S.,
Bouchard, K., Belchior, P., Bottari, C., Par´e, G., Con-
sel, C., and Bier, N. (2020b). Using Ambient Assisted
Living to Monitor Older Adults With Alzheimer Dis-
ease: Single-Case Study to Validate the Monitoring
Report. JMIR Med Inform 2020;8(11):e20215,
Lussier, M., Couture, M., Moreau, M., Lalibert´e, C.,
Giroux, S., Pigot, H., Gaboury, S., Bouchard, K., Bel-
chior, P., Bottari, C., Par´e, G., Consel, C., and Bier,
N. (2020c). Integrating an Ambient Assisted Liv-
ing monitoring system into clinical decision-making
in home care: An embedded case study. Gerontech-
nology, 19(1):77–92.
Organization, W. H. (2015). World report on ageing and
health. World Health Organization.
Pollak, C. P. and Perlick, D. (1991). Sleep Problems and In-
stitutionalization of the Elderly. Journal of Geriatric
Psychiatry and Neurology, 4(4):204–210.
Rashidi, P. and Mihailidis, a. (2013). A Survey on Ambient
Assisted Living Tools for Older Adults. Biomedical
and Health Informatics, IEEE Journal of, PP(99):1.
Robinson, S. M. (2018). Improving nutrition to support
healthy ageing: what are the opportunities for in-
tervention? Proceedings of the Nutrition Society,
Roy, N., Dub´e, R., Despr´es, C., Freitas, A., and L´egar´e, F.
(2018). Choosing between staying at home or mov-
ing: A systematic review of factors influencing hous-
ing decisions among frail older adults. PloS one,
Roy, N., Misra, A., and Cook, D. (2016). Ambient and
smartphone sensor assisted ADL recognition in multi-
inhabitant smart environments. Journal of Ambient
Intelligence and Humanized Computing, 7(1).
Schilit, B., Adams, N., and Want, R. (1994). Context-aware
computing applications. In 1994 First Workshop on
Mobile Computing Systems and Applications, pages
Warren, J. and Marz, N. (2015). Big Data: Principles and
best practices of scalable realtime data systems. Si-
mon and Schuster.
Wood, A. D., Stankovic, J. A., Virone, G., Selavo, L., He,
Z., Cao, Q., Doan, T., Wu, Y., Fang, L., and Stoleru,
R. (2008). Context-aware wireless sensor networks
for assisted living and residential monitoring. IEEE
Network, 22(4):26–33.
Zaharia, M., Chowdhury, M., Franklin, M. J., Shenker, S.,
Stoica, I., and others (2010). Spark: Cluster comput-
ing with working sets. HotCloud, 10(10-10):95.
Zaric, N., Radonjic, M., Pejanovic-Djurisic, M., and
Radusinovic, I. (2015). An example of monitoring
system with reasoning module for ambient assisted
living applications. In IEEE EUROCON 2015 - In-
ternational Conference on Computer as a Tool (EU-
ROCON), number 610658, pages 1–6. IEEE.
Smart CommuniCare 2022 - Special Session on Smart Living Environments to Support Aging-in-Place in Vulnerable Older Adults