ArticlePDF Available

Smart home simulation model for synthetic sensor datasets generation

Authors:

Abstract and Figures

World population is ageing due to longer life expectancy worldwide. There is a trend in elderly people to live alone in their habitual residences in spite of health and safety risks. Smart Homes, intelligent environment systems deployed at elderly homes can act as early warning systems trying to forecast the worsening or exacerbation of the resident chronic conditions. Access to sensor datasets is essential for the development of an efficient real smart home. Procurement of such datasets is subject to several restrictions and difficulties. This paper describes the generation of synthetic datasets by means of a simulation model as a suitable alternative previous to the deployment of a real monitoring system. The collection of synthetic datasets will be used during the next project step to train and evaluate activity recognition methods and algorithms.
Content may be subject to copyright.
7171
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016). Smart home simulation model for synthetic sensor
datasets generation. Sistemas & Telemática, 14(39), 71-84. doi:10.18046/syt.v14i39.2350
Original research / Artículo original / Pesquisa original - Tipo 1
Smart home simulation model for synthetic
sensor datasets generation
Ms.Sc. Darío Weitz / dar.wtz@gmail.com
Denis María / demaria@gmail.com
Franco Lianza / lianza.fl@gmail.com
Nicole Schmidt / nicole.schmidt94@gmail.com
Juan Pablo Nant / jpnant@gmail.com
Departamento Ingeniería en Sistemas de Información / Universidad Tecnológica Nacional, Rosario-Argentina
ABSTRACT World population is ageing due to longer life expectancy worldwide. There is a trend in elderly people to
live alone in their habitual residences in spite of health and safety risks. Smart Homes, intelligent environment systems
deployed at elderly homes can act as early warning systems trying to forecast the worsening or exacerbation of the
resident chronic conditions. Access to sensor datasets is essential for the development of an ecient real smart home.
Procurement of such datasets is subject to several restrictions and diculties. This paper describes the generation of
synthetic datasets by means of a simulation model as a suitable alternative previous to the deployment of a real moni-
toring system. The collection of synthetic datasets will be used during the next project step to train and evaluate activity
recognition methods and algorithms.
KEYWORDS Smart home; intelligent environment systems; sensors; simulation; elderly people.
Modelo de simulación de un hogar inte-
ligente para la generación de datos sin-
téticos de sensores
RESUMEN En el mundo se está verificando un aumento
progresivo del porcentaje de personas mayores en la población
mundial. Hay una tendencia en los adultos mayores a envejecer
en su lugar habitual de residencia, en lugar de utilizar casas de
retiro, a pesar de los riesgos y peligros involucrados. Los hoga-
res inteligentes –sistemas de inteligencia ambiental desplegados
en las residencias de adultos mayores–, pueden brindar apoyo
para compensar el deterioro cognitivo, sensorial o físico de los
mismos. La instalación y puesta a punto de hogares inteligentes
requiere de conjuntos de datos de sensores que sean represen-
tativos de los ambientes que se pretenden monitorear. Existen
diversas restricciones y dificultades al momento de evaluar e im-
plementar hogares inteligentes. Se describe un modelo de simu-
lación que permite la obtención de conjuntos de datos sintéticos
de sensores para ser utilizados en el desarrollo y entrenamiento
de métodos y algoritmos de reconocimiento de actividades, con
el fin de identificar automáticamente cambios en las actividades
de la vida diaria que sugieran deterioros físicos y/o cognitivos.
PALABRAS CLAVE Hogar inteligente; sistemas de inteligen-
cia ambiental; sensores; simulación; adultos mayores.
Modelo de simulação de uma casa inte-
ligente para a geração de dados sintéti-
cos de sensores
RESUMO Está se verificando no mundo um aumento progres-
sivo da percentagem de idosos na população. Há uma tendência
para eles ficarem em seu local de residência habitual, em vez
de usar casas de repouso, apesar dos riscos e perigos envolvi-
dos. As casas inteligentes – sistemas de inteligência ambiental
implantados nos lares de idosos – podem fornecer suporte para
compensar a deterioração cognitiva, sensorial ou física dos mes-
mos. A instalação e funcionamento de casas inteligentes reque-
rem conjuntos de dados de sensores que sejam representativos
dos ambientes que pretendem se monitorar. Existem diversas
restrições e dificuldades na hora de avaliar e programar casas
inteligentes. É descrito um modelo de simulação que permite a
obtenção de conjuntos de dados sintéticos de sensores para ser
usados no desenvolvimento e formação de métodos e algoritmos
de reconhecimento de atividades, a fim de identificar automati-
camente as alterações nas atividades da vida diária que sugiram
deterioração física e/ou cognitiva.
PALAVRAS-CHAVE Casa inteligente; sistemas de inteligên-
cia ambiental; sensores; simulação; idosos.
Received / Recepción: December 2, 2016 - Accepted / Aceptación: December 19, 2016 doi:10.18046/syt.v14i39.2350
72 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
I. Introduction
The progressive increase of the older adults proportion over
the total population is a phenomenon observed in many
countries. The occurrence speed of this phenomenon de-
pends on the living standards and the development degree
of each country, due to for the more developed countries
this transition has been evident for approximately twenty
years. In United States, between 2000 and 2010, the total
population grew 9.7% (from 281.4 million to 308.7 mil-
lion); however, over the same decade the population over
65 years grew to 15.1% (Werner, 2011). It is a clear change
in the trend until the year 2000, where the growth of the
elderly population was slow compared to total growth. In
some European Union [EU] countries with a lower level of
development - Bulgaria, Estonia, Latvia, Lithuania, Hun-
gary and Romania - life expectancy will increase by more
than ten years for men and about eight years for women by
2060 (European Commission, 2014). Thus the more devel-
oped EU states will be reached. The age of the population
will tend to converge between men and women and to get
stabilization country by country, indicating that the transi-
tion towards aging societies is unavoidable worldwide.
The population of Argentine Republic will experience
significant changes between the years 1990 and 2025. It
will increase from 33 million to 47 million, and the mortal-
ity rate will decrease in a context of constant improvement
of health levels, which will be reflected in an increase of 72
to 78 years in the average life expectancy (Torrado, 2004).
These changes, joined to a birth rate reduction, can pro-
duce a progressive increase of the older people percentage
in the total population.
Aging causes a general deterioration of the individual’s
abilities, not only the sensory (sight, hearing, etc.) and mus-
cular capacity, it also alters the cognitive or mental capacity.
Despite the above, older people have presented a trend to-
wards “on site aging”, an opposed lifestyle to the known “re-
tirement homes” or “geriatric residences”. In other words,
older adults prefer the place where they have lived most
of their life, which is familiar and provides them security
in spite of their health issues. So, they decide to spend the
last years of life autonomously in their own home. Hence,
in addition to the environment modification according to
their needs, this situation entrusts the satisfaction of an in-
dependent and autonomous life.
On-site aging cause risks and hazards, such as falls, acci-
dents with appliances, gas network losses or wrong medica-
tion intake. However, even though close relatives of older
I. Introducción
El aumento progresivo de la proporción de adultos mayores
sobre el total de la población es un fenómeno que se obser-
va en numerosos países. La velocidad con que se presenta
dicho fenómeno está en relación con los niveles de vida y
con el grado de desarrollo de cada país, para los más de-
sarrollados esta transición se está evidenciando desde hace
aproximadamente veinte años, en Estados Unidos, en el pe-
ríodo 2000-2010, el total de la población creció 9,7% (pasó
de 281.4 millones a 308.7 millones), sin embargo, durante
la misma década la población mayor a 65 años creció al
15,1% (Werner, 2011). Es un claro cambio en la tendencia
que se verificaba hasta el año 2000, donde el crecimiento de
la población de edad avanzada era lento con respecto al cre-
cimiento total. En algunos países de la Unión Europea [UE]
con menor nivel de desarrollo –Bulgaria, Estonia, Letonia,
Lituania, Hungría y Rumania– la esperanza de vida aumen-
tará más de diez años para los hombres y alrededor de ocho
para las mujeres hasta 2060 (European Commission, 2014).
De esta forma alcanzarán a los Estados más desarrollados
de la UE, de modo que la edad de la población tenderá a
converger entre hombres y mujeres y a estabilizarse país a
país, lo que indica que la transición hacia sociedades enve-
jecidas es inevitable a nivel mundial.
En la República Argentina el panorama demográfico ex-
perimentará cambios significativos entre 1990 y 2025, la
población crecerá de 33 millones a 47 millones, y la tasa de
mortalidad disminuirá en un contexto de mejora constante
de los niveles de salud, lo que se verá reflejado en un aumen-
to de 72 a 78 años en la esperanza de vida promedio (To-
rrado, 2004). Estos cambios, unidos a una disminución de la
tasa de natalidad, se traducirán en un aumento progresivo
del porcentaje de personas mayores en la población total.
Envejecer trae aparejado un deterioro general de las ca-
pacidades del individuo, no solo en lo que respecta a lo
sensorial (vista, audición, etc.) y muscular, sino también en
cuanto a capacidad cognitiva o mental. A pesar de ello, se
está dando la tendencia en los adultos mayores hacia lo que
se denomina “envejecimiento en el lugar”, estilo de vida en
contraposición a los conocidos “hogares de retiro” o “resi-
dencias geriátricas”, que eran habituales hasta hace algunos
años. Aquí el individuo pone en valor el hecho de que el lu-
gar donde ha transitado gran parte de su vida le es familiar
y le da seguridad pese a sus dificultades, por lo que prefiere
pasar sus últimos años de vida de manera autónoma en su
propio hogar. Esta situación, además de permitirle mani-
pular el ambiente según su gusto o necesidad, le confiere la
satisfacción de una vida independiente y autónoma.
Envejecer en el lugar trae consigo riesgos y peligros, tales
como caídas, accidentes con electrodomésticos, pérdidas en
la red de gas o ingesta equivocada de medicamentos. Sin
embargo, aun cuando los familiares cercanos de los adultos
mayores y sus médicos de confianza conocen los mencio-
nados peligros, y a pesar de que es relativamente común la
ocurrencia de accidentes domésticos en los hogares de los
73
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
adults and their physicians are aware of the aforemen-
tioned hazards, and despite the common occurrence of do-
mestic accidents in elderly homes, the on-site aging practice
is very popular. Hence the importance of providing them
a safe environment for the autonomous and independent
development of their daily lives.
Everyday Activities [EA] are parameters used to estimate
the autonomy and independence level of an individual.
They are tasks developed by humans in a day-to-day man-
ner, whose non-implementation implies a greater or lesser
degree of functionality loss (disability). Thus, they have to
depend on third parties (Katz, Ford, Moskowitz, Jackson &
Jae, 1963). There are several physicians specialized in geri-
atrics and gerontology who consider the identification of
changes in EA as one of the ways to detect the appearance
of diseases in their early stages.
Horgas, Wilms and Baltes (1998) classified EA as shown
below: basic, instrumental and advanced. The basic activ-
ities are related to self-care and mobility: sleeping, eating
breakfast, using the bathroom for hygiene, moving around
the property, taking a walk. The instrumental activities are
related to the interaction between humans and objects:
watching television, using a computer, talking on the phone,
using appliances to cook or doing general cleaning; the ad-
vanced activities present less emphasis on the elderly daily
life; these latter are related to work and leisure.
Environmental intelligence systems are composed of
physical components and software entities. They automat-
ically track a certain number of EA. Physical components
include sensors and network systems in order to obtain lo-
cation and activity records: the sensors that allow a location
monitoring or residents activity over time. Its purpose con-
sists of detecting abnormal situations that may be the result
of activity lack or activities performed outside the usual. In
the elderly case, some examples are shown as prolonged
stays in bathrooms or in bed at unusual times.
Environmental intelligence systems deployed in senior
residences can provide support to compensate the cognitive,
sensory or physical impairment through the habitat modifi-
cation, in order to generate information for the resident and
share it with their relatives and professionals who provide
them Health services. They are called Smart Homes [SH],
which have undergone to extensive academic and private
research, such as the Massachusetts Institute of Technol-
ogy (MIT) (2015) “PlaceLab” project; The Florida Gator
Tech Smart House project (2015); The Washington State
University’s Center for Advanced Studies in Adaptive Sys-
mayores, la práctica del envejecimiento en el lugar se da
cada vez con mayor frecuencia, dando origen a la proble-
mática de brindarles un ambiente seguro para el desarrollo
autónomo e independiente de su vida diaria.
Las Actividades de la Vida Diaria [AVD] son parámetros
que se utilizan para estimar el nivel de autonomía e inde-
pendencia de un individuo, son aquellas tareas que los seres
humanos desarrollan de manera cotidiana, cuya no realiza-
ción supone mayor o menor grado de pérdida de funciona-
lidad (discapacidad), lo que conlleva a una dependencia de
terceras personas (Katz, Ford, Moskowitz, Jackson & Jae,
1963). Son numerosos los médicos especializados en geria-
tría y gerontología que consideran a la identificación de
cambios en las AVD como una de las formas para detectar
la aparición de enfermedades en sus primeras fases.
Horgas, Wilms y Baltes (1998) clasificaron a las AVD en:
básicas, instrumentales y avanzadas. Las primeras se rela-
cionan con el auto-cuidado y la movilidad: dormir, desa-
yunar, cenar, utilizar el baño para higiene, moverse por la
propiedad, dar un paseo; Las instrumentales se relacionan
con la interacción entre humanos y objetos: mirar televi-
sión, usar una computadora, hablar por teléfono, utilizar
electrodomésticos para cocinar o hacer la limpieza general
del hábitat; las avanzadas tienen menor énfasis en la vida
cotidiana de los adultos mayores y están relacionadas con el
trabajo y el tiempo libre.
Los sistemas de inteligencia ambiental, conformados por
componentes físicos y entidades de software, permiten ha-
cer un seguimiento automático de un cierto número de
AVD. Los componentes físicos incluyen sensores y sistemas
en red para obtener registros de localización y actividad:
sensores que permiten monitorear la ubicación o actividad
de los residentes a lo largo del tiempo, su objetivo es de-
tectar situaciones anómalas que pueden ser el resultado de
la falta de actividad o de actividades realizadas por fuera
de lo habitual. En el caso de los adultos mayores, algunos
ejemplos son: estadías prolongadas en baños o en la cama
en horarios no habituales.
Los sistemas de inteligencia ambiental desplegados en las
residencias de adultos mayores pueden brindar apoyo para
compensar el deterioro cognitivo, sensorial o físico, trans-
formando el hábitat a fin de generar información del re-
sidente y compartirla con éste, con sus familiares y con los
profesionales que le brindan servicios de salud. Se los deno-
mina Hogares Inteligentes y han sido objeto de profundas
investigaciones académicas y privadas, tales como: el pro-
yecto “PlaceLab” del Massachusetts Institute of Technology
[MIT] (2015); el proyecto “Gator Tech Smart House” de
Florida University (2015); el proyecto Center for Advanced
Studies in Adaptive Systems [CASAS] de la Washington
State University [WSU] (2015); el Phillips Home Lab de la
empresa holandesa Phillips Enterprise Telehealth (2015); y
el Samsung SmartThings, perteneciente a la empresa Sam-
sung (2015). Entre las conclusiones arribadas se indica que
la instalación y puesta a punto de hogares inteligentes es
74 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
tems [CASAS] (2015); The Phillips Home Lab of Phillips
Enterprise Telehealth (2015); And the Samsung Smart-
Things, owned by Samsung (2015). As one of the main
conclusions, the installation and setup of Smart Homes
is a highly complex process, that the industry lacks an
appropriate integration between sensing, data collection,
and remote transmission technologies (Brownsell, Black-
burn & Hawley, 2008). Besides, the idea of remote surveil-
lance as intrusive in the privacy of people with functional
limitations is criticized. Finally, the high cost involved in
transforming a “standard” environment into a “smart”
environment is questioned.
In spite of the above, it is essential to advance in the
smart homes development, since the worsening or exac-
erbation of chronic diseases can be predicted with the
respective decrease of the amount and duration of the
older adults in health care centers. The above can result
in improvements in their life quality, costs of public health
services reduction and the tranquility of the families in-
volved.
The installation and set up of Smart Homes requires
data sets of sensors representative of the environments
that have to be monitored. There are public data sets that
could be used for the training and evaluation of environ-
mental intelligence systems, such as the aforementioned
CASAS project or the data set produced by Van Kasteren
(2015). The latter published the records obtained along 28
days. It included motion sensors, filming, accelerometers
and RFID readers. Its use in other projects is questioned
due to the recruited volunteers did not always belong to
the age group of older adults. Filming is considered an ex-
cessively intrusive monitoring alternative, the data stored
are very specific to the culture of the registration country
and scarcely reliable, because the subjects under study
tend to behave on a dierent way when they are observed.
The implementation and evaluation of a real envi-
ronmental intelligence system describes several con-
straints and diculties: the high cost, due to the need to
experiment with dierent technologies and diverse con-
figurations of sensors and architectures for a long time.
Otherwise, the need to generate a clinical trial protocol
is shown, since experiments are being conducted with
humans. Also, the diculty of recruiting participants,
who must accept the installation of an intrusive system
in their home without (usually) receiving economic com-
pensation. As well as the need to collect data over a very
long period of time, in order to capture those events rep-
un proceso de alta complejidad, que la industria carece de
una apropiada integración entre las tecnologías de sensa-
do, recopilación de datos y transmisión remota (Brownsell,
Blackburn & Hawley, 2008) y se critica la idea de la vigilan-
cia remota como intrusiva en la intimidad de personas con
limitaciones funcionales. Por último, se cuestiona el costo
elevado que implica transformar un ambiente “estándar” en
un ambiente “inteligente”.
Pese a lo anterior, es de fundamental importancia avanzar
en el desarrollo de hogares inteligentes, puesto que los mis-
mos permitirán predecir el agravamiento o exacerbación de
enfermedades crónicas con la correspondiente disminución
en la cantidad y duración de ingresos de los adultos mayores
a los centros de salud. Todo ello redundará en mejoras en
la calidad de vida de los mismos, en la reducción de costos
de los servicios públicos de salud y en la tranquilidad de los
familiares de los mayores involucrados.
La instalación y puesta a punto de hogares inteligentes re-
quiere de conjuntos de datos de sensores que sean represen-
tativos de los ambientes que se pretende monitorear. Existen
conjuntos de datos públicos que podrían ser utilizados para
el entrenamiento y evaluación de sistemas de inteligencia
ambiental, como el ya citado proyecto CASAS o el conjunto
de datos producido por Van Kasteren (2015). Este último
publicó los registros obtenidos durante 28 días con sensores
de movimiento, filmaciones, acelerómetros y lectores RFID.
Su utilización en otros proyectos está cuestionada porque
los voluntarios reclutados no siempre pertenecieron a la
franja etaria de adultos mayores, las filmaciones se conside-
ran una alternativa de monitoreo excesivamente intrusiva,
los datos almacenados son muy específicos de la cultura del
país de registro y escasamente confiables porque los sujetos
bajo estudio suelen comportarse de manera diferente al sen-
tirse observados.
La implementación y evaluación de un sistema real de in-
teligencia ambiental muestra varias restricciones y dificulta-
des: el elevado costo, debido a la necesidad de experimentar
durante un largo tiempo con diferentes tecnologías y diver-
sas configuraciones de sensores y arquitecturas; la necesidad
de conformar un protocolo de ensayo clínico, puesto que se
están realizando experimentos con seres humanos; la difi-
cultad para reclutar participantes, quienes deben aceptar la
instalación en su hogar de un sistema intrusivo sin (habitual-
mente) recibir compensación económica a cambio; la nece-
sidad de recolectar datos durante un período muy prolon-
gado de tiempo, para poder capturar aquellos eventos que
representen conductas típicas de los residentes y eventuales
sucesos anormales; y la dificultad para evaluar escenarios
que impliquen riesgos, tales como caídas, pérdidas de gas,
humos, para los residentes del hogar inteligente (Cardinaux,
Brownsell, Bradley & Hawley, 2013).
A partir de argumentos similares a los indicados previa-
mente, Synnott, Nugent y Jeers (2015) sugieren que la
solución a tales restricciones y dificultades puede consistir
en la generación de conjuntos de datos sintéticos de sen-
75
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
resenting typical residents’ behavior and eventual abnor-
mal events. And finally, the diculty to evaluate scenarios
which involve hazards, such as falls, gas losses, and smoke
for Smart Home residents (Cardinaux, Brownsell, Bradley
& Hawley, 2013).
Synnott, Nugent, and Jeers (2015) suggest that these
constraints and diculties solution may be generated
through synthetic sensor data sets with simulated environ-
mental intelligence systems. The latter facilitates the obtain-
ing of huge data sets, also they allow to experiment and
to set up diverse technological configurations, to plan risks
scenarios and abnormalities at low cost, all above in a rela-
tively short time and without risks or prejudices to eventual
residents. The literature reports a number of simulators or
simulation models that are used to evaluate environmental
intelligence systems, however the high complexity of the
issue has not enabled a definitive solution yet (Paré, Jaana
& Sicotte, 2007; Acampora, Cook, Rashidi, & Vasilakos,
2013).
The objective of the current research consists of describ-
ing an environmental intelligence system simulation model
that generates synthetic data sets of sensors. It will be used
for the development of activity recognition algorithms that
allow to automatically identify changes in the EA that sug-
gest physical and/or cognitive worsening.
II. Simulation
Most of the real world systems such complexity, which
inhibits their resolution through analytical models. For this
reason, the simulation technique is used as an alternative
approach to model the system under study, and to calculate
the performance measures that allows an understanding
of the system functioning. Simulation is defined as the pro-
cess of mathematical model construction or a system or a
decision problem. It is tested through a model, in order to
obtain knowledge or to assist in decision making (Law &
Kelton, 2000). In a particular case of Smart Homes, the
suitable use of the simulation technique allows experiments
to be performed with the purpose of evaluating dierent
settings and obtaining synthetic sensor data in a faster way.
The simulation model described in the current study was
developed using the C # programming language, using Mi-
crosoft Visual Studio 2012 as a development framework.
The modeling of the 3D vision was created by the software
SketchUp pro 2015, using components of the warehouse
3D gallery. Rendering of the model was performed with
SU Plus Podium V2.5 software; Sensor network lines are
sores mediante la utilización de sistemas de inteligencia
ambiental simulados, los mismos facilitan la obtención de
enormes conjuntos de datos, permiten experimentar y po-
ner a punto diversas configuraciones tecnológicas, plantear
escenarios de riesgos y anormalidades a bajo costo, en un
tiempo relativamente breve y sin riesgos o perjuicios a even-
tuales residentes. La literatura reporta un cierto número de
simuladores o modelos de simulación que se emplean para
evaluar sistemas de inteligencia ambiental, pero la elevada
complejidad del problema no ha permitido obtener aún una
solución definitiva (Paré, Jaana & Sicotte, 2007; Acampora,
Cook, Rashidi, & Vasilakos, 2013).
El objetivo del presente trabajo es describir un modelo
de simulación de un sistema de inteligencia ambiental que
genera conjuntos de datos sintéticos de sensores que serán
utilizados para el desarrollo de algoritmos de reconocimien-
to de actividades que permitan identificar automáticamen-
te cambios en las AVD que sugieran deterioros físicos y/o
cognitivos.
II. Simulación
La mayoría de los sistemas del mundo real muestran una
complejidad tal, que inhiben su resolución mediante mo-
delos analíticos. Por tal motivo, se utiliza la técnica de si-
mulación como enfoque alternativo para modelizar al siste-
ma bajo estudio y calcular ciertas medidas de rendimiento
que permiten ganar comprensión sobre el funcionamiento
del mismo. Se define a la simulación como el proceso de
construir un modelo matemático o lógico de un sistema o
problema de decisión, y experimentar con el modelo para
obtener conocimiento del mismo o para asistir en la toma
de decisiones (Law & Kelton, 2000). En el caso particular de
hogares inteligentes, la utilización apropiada de la técnica
de simulación permite realizar experimentos con el objetivo
de evaluar diferentes configuraciones y obtener datos sinté-
ticos de sensores de manera relativamente rápida.
El modelo de simulación descripto en el presente traba-
jo se desarrolló mediante el lenguaje de programación C#,
utilizando Microsoft Visual Studio 2012 como framework
de desarrollo; el modelado de la vista 3D se realizó median-
te el software SketchUp Pro 2015, utilizando componentes
de la galería 3D Warehouse; el renderizado del modelo se
realizó con el software SU Podium V2.5 Plus; las líneas de
redes de sensores se dibujaron mediante Visual Basic Power
Packs Line and Shape Controls; la librería ZedGraph se em-
pleó para la construcción de gráficos y curvas; y el motor de
base de datos para el almacenamiento de datos y resultados
fue Microsoft SQL Server 2008.
El algoritmo de simulación se desarrolló según el esquema
de modelado de simulación de eventos discretos. Es una cla-
se de modelo de simulación que consta de tres dimensiones:
dinámica, discreta y estocástica. La dimensión dinámica
involucra el pasaje del tiempo simulado y una representa-
ción explícita de la secuencia de actividades que realiza el
residente del habitat simulado; la dimensión discreta indica
76 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
drawn using Visual Basic Power Packs Line and Shape
Controls; The ZedGraph library was used for graphs and
curves construction; and finally, the database engine for
storing data and results was Microsoft SQL Server 2008.
The simulation algorithm was developed according to
the discrete event simulation modeling scheme. It is a kind
of simulation model that consists of three dimensions: dy-
namic, discrete and stochastic. The dynamic dimension
involves the passage of simulated time and an explicit
representation of the activities sequence performed by
the resident of the simulated habitat. The discrete dimen-
sion indicates that the events of interest change at sep-
arate points in time. The stochastic dimension allows us
to model the unpredictable behavior of human beings in
their daily activities. The discrete event simulation tech-
nique allows modeling complex systems with a high level
of detail and the resulting model remains “transparent”
for non-technical users.
Discrete event simulation models can be formulated
under three dierent approaches: event-oriented; pro-
cess oriented; or as an activities exploration. The activity
exploration approach emphasizes a review about all the
activities present in the simulation in order to determine
which can start or end the next progress of the simulation
clock. It is computationally less ecient than the other ap-
proaches and therefore it is the least used. However, it is
considered the most appropriate for the elderly daily activ-
ities simulation, due to it allows to describe activities that
occur during fixed intervals of time (day, week, month,
etc.). So, the activity is defined with a pair of events: one
that initiates and another that completes a transformation
operation of the entity state. Activities have finite periods
and state changes are seen at the beginning or end of each
one.
The algorithm includes an animation scheme which al-
lows a visual representation of the sequence of activities
performed by the resident of the simulated habitat. The
animation is defined as a graphical display that modifies its
structure or other property over time and causes the per-
ception of a continuous improvement (Schnotz & Lowe,
2008). Animation contributes to the validation process of
the simulation model - to ensure that the model is an ade-
quate representation of reality - by allowing relatives and/
or health professionals to visualize activities sequence. In
addition, the system includes the possibility of performing
the simulation without the animation process, only when
the objective is the system test or production runs.
que los eventos de interés cambian en puntos separados en
el tiempo; la dimensión estocástica permite modelar la con-
ducta relativamente impredecible que tienen los seres hu-
manos en sus actividades diarias. La técnica de simulación
de eventos discretos permite modelar sistemas complejos
con un elevado nivel de detalle y aún así, el modelo resul-
tante permanece “transparente” para usuarios no técnicos
del mismo.
Los modelos de simulación de eventos discretos pueden
ser formulados bajo tres diferentes enfoques: orientado a
los eventos; orientado a los procesos; o como exploración
de actividades. El enfoque exploración de actividades en-
fatiza una revisión de todas las actividades presentes en la
simulación para determinar cuál puede iniciar o finalizar al
siguiente avance del reloj de la simulación. Es computacio-
nalmente menos eficiente que los otros enfoques y por ello
se lo utiliza menos, pero se considera el más apropiado para
la simulación de las actividades diarias de un adulto mayor
porque permite describir actividades que ocurren durante
intervalos fijos de tiempo (día, semana, mes, etc.). Se define
a la actividad con un par de eventos: uno que inicia y otro
que completa una operación que transforma el estado de
una entidad. Las actividades tienen duraciones finitas y los
cambios de estado se observan al comienzo o al final de la
misma.
El algoritmo incluye un esquema de animación para per-
mitir una representación visual de la secuencia de activida-
des que realiza el residente del habitat simulado. Se define
a la animación como una visualización gráfica que cambia
su estructura u otra propiedad a lo largo del tiempo, y que
dispara la percepción de un cambio continuo (Schnotz &
Lowe, 2008). La animación colabora en el proceso de vali-
dación del modelo de simulación –asegurar que el modelo
sea una adecuada representación de la realidad– al permitir
que los familiares y/o los profesionales de la salud involucra-
dos visualicen la secuencia de actividades. Adicionalmente,
el sistema incluye la posibilidad de realizar la simulación sin
el proceso de animación, cuando el objetivo es la prueba del
sistema o la realización de corridas de producción.
III. SIMULACIÓN DE UN HOGAR INTELIGENTE
Un hogar inteligente es una residencia a la cual se le incor-
poran sensores para monitorear el ambiente, y dispositivos
para proveer servicios proactivos que mejoran la calidad de
vida y seguridad de los residentes. La simulación de hogares
inteligentes para la generación de datos sintéticos se puede
realizar a través de dos enfoques claramente diferenciados:
enfoques basados en modelos o enfoques interactivos (Syn-
nott et al., 2015). El primer enfoque se basa en la especifi-
cación de modelos de actividad que describen el orden de
los sucesos, la probabilidad de su ocurrencia, el tiempo de
duración de la actividad y el rango de valores de los sen-
sores que definen el estado de una entidad. La calidad del
conjunto de datos resultante depende básicamente de la ca-
lidad del modelo de actividad subyacente en el modelo de
77
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
III. Smart home simulation
A Smart Home is a residence compound of sensors for en-
vironment monitoring, and devices to test proactive services
that improve the life quality and the resident´s safety. Smart
Home simulation for synthetic data generation is generated
by clearly dierentiated approaches: model-based approach-
es or interactive approaches (Synnott et al., 2015). The first
approach is based on activity models specification which de-
scribes events order, the occurrence probability, the activity
duration and the senses values range that define the state of
an entity. The quality of the data set resulted basically from
the quality of the underlying activity model in the simula-
tion model, which depends on the appropriate information
availability about the activities performed by the simulated
resident. On the other hand, the interactive approaches em-
phasize the model of the virtual environments and the virtual
sensors by the modeling of the activities. Besides, they facili-
tate the obtaining of synthetic data related to modifications
in the environments or in the quantity, deployment or sensors
location.
The simulation model described in this paper uses the
model-based approach. It is compound by three compo-
nents: the resident of the simulated habitat; the simulated en-
vironment; and the sensor network and system architecture.
The first component shows the model restriction to the EA
detection of a single resident. It does not assume a severe
restriction since the coexistence of two more people in a res-
idence allows a mutual evaluation of the health status which
exceeds any information system.
The simulated environment follows the general guidelines
of Smart Homes described in the introduction. They consist
of a bedroom, a living room with TV, a study with comput-
er, a fully equipped kitchen with a table and chairs, and a
bathroom.
The simulated sensor network corre-
sponds to wireless sensors embedded in
the environment. It is a set of security
devices, physically distributed along the
monitoring residence, communicated
by some type of wireless technology. It
presents a flexible topology and it does
not require a minimum number for its
operation. A sensor management mod-
ule allows the incorporation of several
brands, models, and suppliers of move-
ment sensors, temperature, humidity,
bed occupancy and/or armchair (Fig-
simulación, el cual a su vez depende de la disponibilidad de
información apropiada sobre las actividades que realiza el
residente simulado. Por su parte, los enfoques interactivos
ponen el énfasis en el modelado de los ambientes virtuales y
de los sensores virtuales por sobre el modelado de las activi-
dades, y facilitan la obtención de datos sintéticos relaciona-
dos con cambios o modificaciones en los ambientes o en la
cantidad, despliegue o ubicación de los sensores.
El modelo de simulación descrito en el presente trabajo
utiliza el enfoque basado en modelos y está conformado por
tres componentes: el residente del habitat simulado; el am-
biente simulado; y la red de sensores y la arquitectura del
sistema.
El primer componente muestra que el modelo está restrin-
gido a la detección de las AVD de un único residente, no
implica una severa restricción debido a que la coexistencia
de dos o más personas en una residencia permite una mu-
tua evaluación del estado de salud superior al de cualquier
sistema de información.
El ambiente simulado sigue los lineamientos generales de
los hogares inteligentes descritos en la introducción y consta
de un dormitorio, un living con televisor, un estudio con
computadora, una cocina completamente equipada con
una mesa y sillas, y un baño.
La red de sensores simulada corresponde a sensores ina-
lámbricos embebidos en el ambiente, se trata de un conjun-
to de dispositivos autónomos, distribuidos físicamente en la
residencia a monitorear, que se comunican mediante algún
tipo de tecnología inalámbrica, su topología es flexible y no
requieren de un número mínimo para su funcionamiento.
Un módulo de gestión de sensores permite la incorporación
de diversas marcas, modelos y proveedores de sensores de
movimiento, temperatura, humedad, ocupación de cama
y/o sillón (Figura 1). También sensores magnéticos de
apertura de puertas, de sonido, fotocontrol y detectores de
humo, gas natural y monóxido de carbono. Haciendo clic
derecho sobre cuadrados azules dibujados en los distintos
ambientes de la residencia simulada se pueden agregar o
quitar sensores, activarlos o desactivarlos. El sistema permi-
Figure 1. Sensor’s management module / Módulo de gestión de sensores
78 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
ure 1). As well as magnetic sensors for opening doors, sound,
photo control and smoke detectors, natural gas and carbon
monoxide. Right clicking on blue squares is drawn in dier-
ent environments of the simulated residence. They can add
or remove sensors, activate or deactivate them. This system
allows to setup the probability of a disaster occurrence (fire,
loss of gas, etc.), through the probability distribution between
the dierent rooms of the residence.
The underlying activity model proposes an EA sequence
generated from specific probability distributions. They cor-
respond to the start time and the duration of basic and in-
strumental activities set. Such probability distributions are
obtained from empirical data collection, and followed by a
consistency analysis and the use of data adjustment software.
The simulation proceeds according to a mechanism of four
stages: Generation of the sequence of activities; Advance of
the simulated time to the occurrence time of the next activ-
ity; Activation of the corresponding sensors; and database
storage of simulated time, activity and activated sensors. If
the occurrence time of one or more activities overlaps with
another activity in process, the algorithm places the first in a
queue and executes them according to a scheme FIFO [First
In First Out]. The sequence of random numbers necessary
to generate the random variables of probability distributions
is obtained from the generator provided by programming
language. The microprocessor clock or a value entered from
the keyboard is used as seed. The latter alternative allows the
sequence of random numbers to be used throughout the ver-
ification step in order to ensure that the simulation model is
free of logical errors.
te configurar la probabilidad de ocurrencia de un siniestro
(incendio, pérdida de gas, etc.), distribuyendo esa probabili-
dad entre las diferentes habitaciones de la residencia.
El modelo de actividad subyacente propone una secuencia
de AVD que se genera a partir de distribuciones de proba-
bilidad específicas que corresponden a la hora de inicio y a
la duración de un conjunto de actividades básicas e instru-
mentales. Tales distribuciones de probabilidad se obtienen
a partir de la recolección de datos empíricos, seguido de
un análisis de consistencia y la utilización de software de
ajuste de datos. La simulación procede según un mecanismo
de cuatro etapas: generación de la secuencia de activida-
des; avance del tiempo simulado al tiempo de ocurrencia
de la próxima actividad; activación de los correspondien-
tes sensores; y almacenamiento en base de datos del tiempo
simulado, actividad y sensores activados. Si el tiempo de
ocurrencia de una o más actividades se solapa con otra acti-
vidad en proceso, el algoritmo coloca a las primeras en una
cola de espera y procede a ejecutarlas según un esquema
FIFO [First In First Out]. La secuencia de números alea-
torios imprescindibles para generar las variables aleatorias
correspondientes a las distribuciones de probabilidad se
obtiene a partir del generador provisto por el lenguaje de
programación utilizando como semilla el reloj del micropro-
cesador o un valor ingresado desde teclado. Esta última al-
ternativa permite repetir la secuencia de números aleatorios
y se la utiliza durante la etapa de verificación, para asegurar
que el modelo de simulación esté libre de errores lógicos.
IV. Resultados y discusión
El perfil del residente del habitat simulado corresponde a
un adulto mayor que vive solo en su residencia habitual, po-
see algún tipo de enfermedad crónica, pero no se encuentra
limitado en sus desplazamientos por el interior o exterior
de la propiedad. Los datos para generar la secuencia de ac-
Figure 2. Adjust to activity “breakfast” - duration /Ajuste de la actividad desayunar - duración
79
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
IV. Results and discussion
The resident profile of the simulated habitat corresponds
to an older adult who lives only in his or her habitual resi-
dence, somebody who has some type of chronic illness, but
is not limited in his or her movements inside or outside the
property. In order to generate the sequence of activities,
the data were obtained from the filling of a daily activities
sheet performed by an 81 year old woman living alone in
her private home. Some missing data, by default or forgot-
ten, were completed using the “Random Data” function of
the “Calc” Menu belong to Minitab software. Data were
entered into a MS-Excel spreadsheet and then they were
processed through the Ball software (Evans & Olson, 1998).
It allowed obtaining the probability distribution functions
(with their corresponding parameters) of the start time and
duration of basic and instrumental activities that best fit the
data collected. Figure 2 describes the best fit determined
by the Crystal Ball software to the data collected for the
“Duration” property of the “Breakfast” activity. Taking into
account other activities recorded on the form, the following
probability distributions were obtained - the information
in parentheses indicates the value of the Anderson-Dar-
ling test): sleep: start time: Weibull (0.5227); Breakfast: start
time, beta (0.1635); Duration, negative binomial; Lunch:
start time, logistics (0.3434); Duration, negative binomial;
Snack: start time, normal (0.2004), duration, binomial;
Dinner: start time, normal logarithmic (0.1940), duration,
Poisson.
Wireless sensor networks were tested in residences of
families of the project authors. Sensor performance, in-
stallation process, power consumption and communication
using the ZigBee protocol (IEEE 802.15.4) were evaluated.
Figure 3 shows how the mixed use of temperature, hu-
midity and luminosity sensors installed in a bath allow the
detection of EA “take a shower”.
tividades se obtuvieron a partir del llenado de una planilla
de actividades diarias realizadas por una mujer de 81 años
de edad que vive sola en su domicilio particular. Algunos
datos ausentes, por omisión u olvido, fueron completados
mediante la función “Datos Aleatorios” del Menú “Calc”
del software Minitab. Los datos fueron ingresados a una pla-
nilla en MS-Excel y posteriormente procesados mediante
el software Crystal Ball (Evans & Olson, 1998) para obte-
ner las funciones de distribución de probabilidad (con sus
correspondientes parámetros) de la hora de inicio y dura-
ción de las actividades básicas e instrumentales que mejor
ajustaron a los datos recolectados. La Figu ra 2 muestra el
mejor ajuste determinado por el software Crystal Ball a los
datos recolectados para la propiedad “Duración” de la ac-
tividad “Desayunar”. Para otras actividades registradas en
la planilla se obtuvieron las siguientes distribuciones de pro-
babilidad –la información entre paréntesis indica el valor
de la prueba Anderson–Darling): dormir: hora de inicio:
Weibull (0,5227); desayunar: hora de inicio, beta (0,1635);
duración, binomial negativa; almorzar: hora de inicio, logís-
tica (0,3434); duración, binomial negativa; merendar: hora
de inicio, normal (0,2004), duración, binomial; cenar: hora
de inicio, logarítmico normal (0,1940), duración, Poisson.
Las redes de sensores inalámbricos fueron ensayadas me-
diante su instalación en residencias de familiares de autores
del proyecto. Se evaluó el funcionamiento de los sensores,
la facilidad de instalación, el consumo de energía y la co-
municación mediante el protocolo ZigBee (IEEE 802.15.4).
La Figur a 3 muestra cómo el uso combinado de sensores
de temperatura, humedad y luminosidad instalados en un
baño permite detectar la realización de la AVD “tomar una
ducha”.
Previo al inicio de una corrida de simulación se debe ac-
tivar el conjunto de sensores que caracterizan el escenario
a evaluar. Se define la fecha de inicio y la cantidad de días
a simular, y se elige entre una simulación con animación –
para visualización del escenario o validación del modelo– o
una simulación sin animación –para obtención rápida de
resultados–. Se configura la probabilidad de ocurrencia de
un siniestro con la correspondiente distribución de esa pro-
babilidad entre las diferentes habitaciones de la residencia.
Figure 3. Data collected from sensors in restroom environment / Datos de sensores recolectados en el ambiente “baño”
80 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
Prior to the starting of a simulation run, it must be acti-
vated the set of sensors that characterizes the scenario to
be evaluated. The start date and the number of days for
simulation are defined, and a simulation with animation is
chosen – in pro of scenario visualization or model valida-
tion - or a simulation without animation – in order to ob-
tain faster results. The probability of sinister
occurrence with the respect distribution of
that probability between the dierent rooms
of the residence is configured.
The system creates an activities sequence
for each simulated day. Figure 4 shows the
start time, the end time and the duration of
the activities to be performed by the resi-
dent between 0:00 and 14:03 hours for a
simulated day. The activities have associated
a priority value in order to solve situations
of “tie” in occurrence time. As mentioned
above, wheter the occurrence time of one or
more activities overlaps with another activity
in process, the algorithm places the first ones
in a queue and proceeds to execute them ac-
cording to a FIFO scheme. Some EAs (going
to the bathroom, receiving a phone call, tak-
ing mate) have a break property for the rest
El sistema genera una secuencia de actividades para cada
día simulado. La Fi gura 4 muestra el horario de inicio, el
horario de finalización y la duración de las actividades a
realizar por el residente entre las 0:00 y las 14:03 horas para
un día simulado. Las actividades tienen asociado un valor
de prioridad para resolver situaciones de “empate” en tiem-
po de ocurrencia. Como se indicó, si el tiempo de ocurren-
cia de una o más actividades se solapa con otra actividad en
proceso, el algoritmo coloca a las primeras en una cola de
espera y procede a ejecutarlas según un esquema FIFO. Al-
gunas AVD (ir al baño, recibir una comunicación telefónica,
tomar mate) tienen una propiedad de interrupción del resto
de las demás actividades. Cuando el algoritmo detecta una
interrupción, procede a almacenar el tiempo restante de la
actividad interrumpida y reacomoda la lista para que esa
actividad sea reasumida al finalizar la interrupción. Duran-
te la corrida, la información recolectada por los sensores se
visualiza en un log de sensores y se almacena en una base
de datos. La Figura 5 muestra la pantalla principal del mo-
delo durante una corrida con animación. El log de sensores
indica la transición del residente simulado desde el baño
hacia el living comedor a las 7:36 horas del día simulado 1
de enero de 2016.
La Figura 6 muestra un reporte de métricas donde se indica
el día y la hora del registro, el ícono del sensor que está regis-
trando la información, el ambiente de registro, el valor infor-
mado por el sensor y el modelo de sensor elegido para la co-
rrida. Se almacenaron en la base de datos 1.055.235 registros
durante una corrida de simulación correspondiente a 90 días.
El sistema incluye un conjunto de aplicaciones gráficas que
permite visualizar información puntual e información resumi-
da referida a los AVD del residente monitoreado. Así, la Figu-
ra 7 muestra el porcentaje de ocupación diaria de los distintos
Figure 4. Activities list / Lista de actividades
Figure 5. Main view of simulation model / Vista principal del modelo de simulación
81
ambientes (incluyendo salida al exterior) para un período si-
mulado de quince días. Como ejemplo, en el día 8 se observa
una menor ocupación del dormitorio y una mayor ocupación
del comedor. Si esta tendencia se hubiera prolongado durante
varios días, podría indicar un posible deterioro en la condición
física o cognitiva que justificaría la intervención de un profesio-
nal de la salud asociado. Solo se muestran quince días por una
cuestión de claridad de visualización de la gráfica.
V. Conclusiones
Una de las alternativas más prometedoras para facilitar
el envejecimiento en el lugar habitual de residencia de los
adultos mayores radica en los hogares inteligentes: sistemas
de inteligencia ambiental desplegados en las residencias que
of other activities. Once the algorithm detects an interrupt,
it proceeds to store the remaining time of the interrupted
activity and resets the list so that activity is taken at the end
of the interruption. Throughout the run, the information
collected by the sensors is displayed in a sensor log and
stored in a database. Figure 5 shows the main screen of
the model along an animated run. The sensor log indicates
the transition of the simulated resident from the bathroom
to the living room at 7:36 p.m. on the simulated day Janu-
ary 1st, 2016.
Figure 6 shows a metric report that describes the date
and registration time, the sensor icon that is recording the
information, the recording environment, the value reported
by the sensor and the sensor model chosen for the simula-
tion run. 1,055,235 records were stored in a database with
90-day simulation run.
The system includes a set of graphical applications that
allow the visualization of timely information and summary
information related to the EA of the monitored resident.
Thus, Figure 7 shows the percentage of daily occupancy in
several environments (including outbound) for a simulated
period of fifteen days. As an example, on day 8 there is a
lower occupancy of the bedroom and a greater occupation
of the dining room. If this trend has extended for several
days, it may indicate a possible deterioration in the physi-
Figure 6. Metrics report / Reporte de métricas
Figure 7. Environment’s occupation / Ocupación de ambientes
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
82 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
cal or cognitive condition that requires an associated health
professional intervention. Only fifteen days are shown to
obtain a display understanding of the graph.
V. Conclusions
One of the most promising alternatives to facilitate ag-
ing in elderly residence takes place in Smart Homes: en-
vironmental intelligence systems deployed in residences
that modify habitat with the aim of generating residents’
information and sharing with them, their relatives and pro-
fessionals who provide health services. However, monitor-
ing and remote assistance of older adults with some form
of physical or cognitive impairment is a highly complex
technological problem. Its implementation is expensive and
there are clinical and ethical considerations that must be
taken into account.
In order to reduce the constraints and diculties about
the implementation and evaluation of a real environmen-
tal intelligence system, it is recommended to use simulation
models which generate synthetic sensor data.
In the current research, it was described a simulation
model which allowed the obtainment of the mentioned
synthetic data sets of sensors. These sensors were used for
the development and training of methods and recognition
of activities algorithms with the aim of identifying changes
in EA that suggest physical and/or cognitive impairments.
Hence, it is created an early warning system which im-
proves the elderly life quality and reduces the amount and
duration of the older adults in health care centers.
transforman el hábitat con el objetivo de generar informa-
ción del residente y compartirla con éste, con sus familiares
y con los profesionales que le brindan servicios de salud.
Sin embargo, el monitoreo y la asistencia remota de adultos
mayores con algún tipo de discapacidad física o cognitiva
es un problema tecnológico de elevada complejidad. Su im-
plementación es costosa y existen consideraciones clínicas y
éticas que deben ser tenidas en cuenta.
Para reducir las restricciones y dificultades relacionadas
con la implementación y evaluación de un sistema real de
inteligencia ambiental, se recomienda la utilización de mo-
delos de simulación que permitan generar datos sintéticos
de sensores.
En el presente trabajo, se describe un modelo de simu-
lación que permitió la obtención de los citados conjuntos
de datos sintéticos de sensores. Los mismos serán utilizados
para el desarrollo y entrenamiento de métodos y algorit-
mos de reconocimiento de actividades con el objetivo de
identificar automáticamente cambios en las AVD que su-
gieran deterioros físicos y/o cognitivos. De esta manera, se
obtendrá un sistema de alerta temprana que mejorará la ca-
lidad de vida de los adultos mayores y reducirá la cantidad
y duración de ingresos de los mismos en centros de salud
asistencial.
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
83
References / Referencias
Acampora, G., Cook, D.J., Rashidi, P., & Vasilakos, A.V. (2013). A survey on ambient intelligence in healthcare. Proceedings
of the IEEE, 101 (12), 2470-2488.
Brownsell, S., Blackburn, S., & Hawley, M.S. (2008). An evaluation of second and third generation telecare services in older
people’s housing. J. Telemed Telecare, 14 (1), 8-12.
Cardinaux, F., Brownsell, S., Bradley, D. & Hawley, M.S. (2013). A home daily activity simulation model for the evaluation of
lifestyle monitoring systems. Computers in Biology and Medicine, 43, 1428-1436.
CASAS, Center for Advanced Studies in Adaptive Systems. Retrieved from http://casas.wsu.edu/research-projects
European Commission. (2014). The 2015 Ageing Report. European Economy, (8), 11.
Evans, J.R. & Olson, D.L. (1998). Introduction to simulation and risk analysis. New Jersey, NJ: Prentice Hall.
Horgas, A., Wilms, H., & Baltes, M. (1998). Daily life in very old age: Everyday activities as expression of successful aging.
The Gerontologist, 38(5), 556-567.
Katz, S., Ford, A.B., Moskowitz, R.W., Jackson, B.A., & Jaffe, M.W. (1963). Studies of illness in the aged: The index of ADL,
a standardized measure of biological and psychosocial function. JAMA, 185, 914-919.
Law, A.M. & Kelton, W.D. (2000). Simulation modeling and analysis [3th ed.]. Singapore: McGraw-Hill
Massachusetts Institute of Technology [MIT]. (2015). Media Lab Project April 2015. Cambridge, MA: MIT. Available at: http://
www.media.mit.edu/files/projects.pdf
Mobile and Pervasive Computing Research, University of Florida. Retrieved from http://www.icta.ufl.edu/gt.htm
Paré, G., Jaana, M., & Sicotte, C. (2007). Systematic review of home telemonitoring for chronic diseases: The evidence base.
Journal of the American Medical Informatics Association, 14(3), 269 -275.
Phillips Enterprise Telehealth. Retrieved from http://www.usa.philips.com/healthcare/solutions/enterprise-telehealth
Samsung SmartThings. Retrieved from https://www.smartthings.com/uk/
Schnotz, W. & Lowe, R.K. (2008). A unified view of learning from animated and static graphics. In: Lowe, R.K., Schnotz, W.
(Eds.), Learning with animation: Research implications for design (pp. 304-356). New York, NY: Cambridge University
Press.
Synnott, J., Nugent, C., & Jeffers, P. (2015). Simulation of smart home activity datasets. Sensors, 15, 14162-14179.
Torrado, S. (2004). Argentina: escenarios demográficos hacia 2025 [Informe preparado para el Programa de Estudios Pro-
spectivos de la Secretaría para la Ciencia, la Tecnología y la Innovación Productiva (SECTIP)]. Buenos Aires, Argentina:
UBA. Available at: http://www.econ.uba.ar/www/departamentos/economia/nuevo/depto/materias_depto/cursos/557_gar-
ciadefanelli/Programa%202009/Torrado%20Escenarios.pdf
Van Kasteren, T. Datasets for activity recognition. Retrieved from: https://sites.google.com/ site/tim0306/datasets
Werner, C. (2011, Nov.). The older population: 2010 [2010 Census Briefs]. Washington, DC: United States Census Bureau.
Available at: https://www.census.gov/prod/cen2010/briefs/c2010br-09.pdf
Smart home simulation model for synthetic sensor datasets generation. Sistemas & Telemática, 14(39), 71-84
84 http://www.icesi.edu.co/revistas/index.php/sistemas_telematica
CURRICULUM VITAE
Darío Weitz Chemical Engineer, Master in International Relationships. Associated professor at the Facultad
Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentina, in “Control theory” and “Simulation”.
Project Director in “Sensors and systems for environments that improve monitoring and remote assistance to
older people” - Code: PID UTN 3784 / Ingeniero Químico con Maestría en Relaciones Internacionales; profesor
asociado Teoría de Control, Facultad Regional Rosario, Universidad Tecnológica Nacional, Argentina; profesor
titular Simulación, Facultad Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentina. Director
del Proyecto de R&D “Sensores y sistemas para ambientes que faciliten el monitoreo y la asistencia remota de
adultos mayores” - Código: PID UTN 3784.
Denis María Last year student of Systems and Informatics Engineering at Facultad Regional Rosario, Universi-
dad Tecnológica Nacional, Rosario, Argentina / Alumno del último año de la carrera de Ingeniería en Sistemas de
Información, Facultad Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentina.
Franco Lianza Last year student of Systems and Informatics Engineering at Facultad Regional Rosario, Univer-
sidad Tecnológica Nacional, Rosario, Argentina / Alumno del último año de la carrera de Ingeniería en Sistemas
de Información, Facultad Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentina.
Nicole Schmidt Last year student of Systems and Informatics Engineering at Facultad Regional Rosario,
Universidad Tecnológica Nacional, Rosario, Argentina / Alumno del último año de la carrera de Ingeniería en
Sistemas de Información, Facultad Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentina.
Juan Pablo Nant Last year student of Systems and Informatics Engineering at Facultad Regional Rosario,
Universidad Tecnológica Nacional, Rosario, Argentina / Alumno del último año de la carrera de Ingeniería en
Sistemas de Información, Facultad Regional Rosario, Universidad Tecnológica Nacional, Rosario, Argentinaa.
Weitz, D., María, D., Lianza, F., Schmidt, N., & Nant, J.P. (2016).
... The table shows that most of the papers did not consider modelling actuators. The reason is that the applications of the simulators are mostly either dataset generation for quickly designing and testing activity recognition models, especially in residential buildings (such as [15], [29]) or quickly evaluating different sensor deployment configurations (such as [30], [63]). In addition, most of the papers used an interactive approach for agent modelling because the approach principally offers more variation of agent traces than model-driven approaches if a large enough number of users interact with the simulators. ...
... Weitz et al. [63] developed a simulator to generate synthetic datasets suitable for training and testing machine learning models for recognizing ADL activities. The simulator uses the discrete event simulation modelling approach, operating on three dimensions: dynamic, discrete and stochastic. ...
Article
Research on smart indoor space applications, which focus on the occupants’ trace, requires a design process consisting of deploying several devices (sensors and actuators) and developing software components (services). The design process in the real-world is labour-intensive, costly and time-consuming, which implies the need for a simulation-based prototyping framework for quickly simulating smart indoor space applications. An increasing body of literature has sought to provide simulation methodologies for smart indoor space applications. However, a unified taxonomy that represents the simulation tools in terms of their essential components and their specifications is missing. In this survey paper, we systematically collect the related papers and derive a taxonomy for the specifications of the smart indoor space simulation tools. We use the taxonomy in order to cluster the papers. Our results indicate that our proposed taxonomy can divide papers into meaningful groups. We review each group in detail, summarize each paper, and provide insights into their specifications.
... Smart interfaces allow the user to communicate with the devices in the smart environment in a simple way (Weitz, D. et al. 2016). The main goal of these interfaces is to hide the complexity of the system from the user and show only its functionalities. ...
Chapter
Full-text available
Vision Science understands that vision is a multisensory, perceptual and kinesthetic process and/or with the ability to process information from the environment, obtain meaning and understand what is seen through the visual system”. The retina is the structural organ of vision with a photosensitive membrane, which contains photoreceptors: the cones that are responsible for day vision (photopic vision) and the rods responsible for night vision (scotopic vision). (Vicario, C.U. 1999) (Fransoy & Augé M., 2013). The trichromatic theory argues that the retina contains three types of particles which select and produce color effects and color discrimination results from the mixture of product colors. The three receptors mainly respond to red, yellow and blue light. The theory of the opposite color deriving from the subjectivity of color and the understanding of color vision was achieved when it was discovered that visual perception is processed in a series of zones (Chaparro Morales, I. 2018)
... Smart interfaces allow the user to communicate with the devices in the smart environment in a simple way (Weitz, D. et al. 2016). The main goal of these interfaces is to hide the complexity of the system from the user and show only its functionalities. ...
Chapter
As an integral part of Industry 4.0, the intelligent humidifier based on mechatronic control has a strong potential for both industrial and artisanal applications. This project is mainly based on the application of mechatronic engineering, where the branches of mechanical electronics and systems are implemented for the control of a device. This project intends to help and improve the quality of life of the population of Ciudad Juarez, Chihuahua, Mexico, through innovative technology and smart service. It is intended to apply a level control by programming a PID Fuzzy controller, using LabVIEWbased on IoT in a Smart city.
... Smart interfaces allow the user to communicate with the devices in the smart environment in a simple way (Weitz, D. et al. 2016). The main goal of these interfaces is to hide the complexity of the system from the user and show only its functionalities. ...
Chapter
According to data obtained from the National Institute of Statistics and Geography (INEGI) in the state of Chihuahua, up to 35% of people with disabilities have difficulty moving or using their arms (INEGI, 2017). The cause of their impairments may differ but ultimately causing them not to perform effectively their everyday activities due to lack of strength. This is why rehabilitation is used as a way to help improve or recover the movement of the upper limbs. The scientific community has shown interest in improving the way therapies are performed by using exoskeletons. There is vast research on the use of exoskeletons, most of which were designed to maintain a rigid posture centered on the back, legs and hips. Thus, it is of great interest to improve the design, increase comfort and reduce the effects caused by this condition, to provide strength to the upper extremities and thereby perform mundane activities without discomfort such as moving a box or putting a book on a shelf.
... Smart interfaces allow the user to communicate with the devices in the smart environment in a simple way (Weitz, D. et al. 2016). The main goal of these interfaces is to hide the complexity of the system from the user and show only its functionalities. ...
Book
Full-text available
Reliability engineering isba powerfull tool in industry 4.0 in perder to predict real life expectation of software and hardware which help in te dessition making procces
... This growth causes an imbalance in the costs of public health services and the probable shortage of direct care workers such as nurses, paramedics and health aides. In addition, there are a significant number of older adults with progressive health problems who prefer to live in their homes where they have been established for many years [7]. ...
Chapter
Over the years, people tend to fail to recall different activities of everyday life, especially those that are performed less frequently. Forgetting a medical appointment, a family member’s birthday, or to take a medicine, is a common problem for many older adults. Although Information and Communication Technologies provide several options to help older adults remember activities like these, many of them could feel discouraged by not being able to properly use these tools and take full advantage of them. In other cases, elderly people may feel intimidated and even refuse to interact with technological devices. For these reasons, this paper proposes the use of a conventional object, which can be found in any home, to help older adults remember certain activities. Specifically, an ordinary chair has been selected to be employed as a device to provide the necessary reminders. The reminders implemented in the designed prototype are provided by audio, lights and vibrations, things that users do not notice at first sight the presence of technology in the chair. This prototype was evaluated through a user study with the collaboration of older adults. The results of the evaluation were positive, which concludes that the proposal has a favorable reception. Thus, this proposal could provide an important contribution to the major goal of helping to improve the quality of life of the elderly population.
Article
Full-text available
Recent advances in technology have provided a wide variety of possibilities for incorporating animation in computer-based learning environments. In a very general sense, the term ‘animation’ can refer to any display element that changes its attributes over time. Taken this broadly, examples such as words that fly across the screen or objects that vibrate, blink, or change their color would be regarded as animations (Wright, Milroy, & Lickorish, 1999). In this chapter, we use the term in a more restricted sense. We define ‘animation’ as a pictorial display that changes its structure or other properties over time and which triggers the perception of a continuous change. Our definition includes examples of dynamic visualization such as pictorial displays that present objects continuously from different perspectives, show the assembly of a complex object from its parts, present the functioning of a technical device such as a bicycle pump, display the dynamic behavior of a meteorological system, or model the co-variation of variables in a graph (Bodemer, Ploetzner, Feuerlein, & Spada, 2004; Hegarty, Narayanan, & Freitas, 2002; Lowe, 2004; Mayer, 2001, 2005; Mayer & Moreno, 2002; Schwan & Riempp, 2004). Although the use of animation has been enabled and stimulated by technology, our definition excludes technical considerations such as the number of frames per second and whether or not the pictorial display is computer generated. However, our definition does include video (cf. Baek & Layne, 1988; Betrancourt & Tversky, 2000). Steven Spielberg’s film "Jurassic Park" demonstrated that viewers cannot distinguish sequences generated by computer from those filmed in a conventional studio. Although these differences in how sequences were generated may be important from a technical perspective, they are obviously of no psychological relevance to the viewer. On the one hand, animations are frequently considered as inherently superior to static pictures. Whereas static pictures display only visuospatial information, animations display temporal information as well. Accordingly, animations can be considered as more informative, more natural or more ‘authentic’ than the corresponding static pictures. Thus, it is very tempting to assume that learning from animation will be more effective than learning from static pictures. On the other hand, learning from animation can also be seen as more demanding because more information has to be processed and this imposes a higher cognitive load on the learner’s working memory (Sweller, van Merriënboer, & Paas, 1998). These two views encapsulate the prevailing tendency to emphasize differences between animated and static graphics rather than to consider their commonalities. From a technological perspective, there may well be clear differences between static and animated pictures with regard to the physical nature of the representation. However, from a psychological perspective, they are fundamentally linked in terms of the human information processing capacities that operate upon them both during learning. Our evolutionary history has given us a perceptual and cognitive system that is well equipped to cope with world in which change is part of the normal condition. The highly dynamic environment in which we live usually includes both temporally stable and temporally changing components. In our everyday lives, we do not continually compartmentalize our environment into static and dynamic parts. Rather, we deal with these components in an integrated and flexible manner as we continually construct a coherent functional mental representation of the world around us. Even when we move through a static environment, our visual field is continuously changing. The flux of our everyday experience is thus a seamless blend of changing and unchanging stimuli without fixed boundaries between static and dynamic components. On this basis, it seems difficult to justify a sharp distinction between learning from animated and learning from static pictures, and there appears to be little reason to assume that animations are necessarily easier or more demanding than static pictures. In this chapter, we will therefore argue for a more unified approach to dealing with animations and static pictures. The basis for this argument is that animations and static pictures are not fundamentally different from a psychological point of view because animated and static pictures are processed by the same perceptual and cognitive system. Given that this system has evolved by interacting with a complex environment containing an ongoing blending of both dynamic and stable stimuli, the same set of basic perceptual and cognitive principles should underlie learning from both types of representation. It follows that the principles for using animation effectively are not fundamentally different from those for using static pictures and that the former can be considered as an extension of the latter. With regard to educational practice, we suggest that designers and developers of learning resources should not consider animation as something totally different in a qualitative sense and that therefore requires totally new design principles. Rather, they could learn from our history’s rich legacy of visual communication in order to understand and apply underlying principles that are common to static and animated graphics. With regard to educational theory, we suggest that researchers should not ‘re-invent the wheel’ by considering learning from animation as a fundamentally different branch of scientific analysis. Rather, they should view this topic as closely related to other research on visual communication.
Article
Full-text available
A globally ageing population is resulting in an increased prevalence of chronic conditions which affect older adults. Such conditions require long-term care and management to maximize quality of life, placing an increasing strain on healthcare resources. Intelligent environments such as smart homes facilitate long-term monitoring of activities in the home through the use of sensor technology. Access to sensor datasets is necessary for the development of novel activity monitoring and recognition approaches. Access to such datasets is limited due to issues such as sensor cost, availability and deployment time. The use of simulated environments and sensors may address these issues and facilitate the generation of comprehensive datasets. This paper provides a review of existing approaches for the generation of simulated smart home activity datasets, including model-based approaches and interactive approaches which implement virtual sensors, environments and avatars. The paper also provides recommendation for future work in intelligent environment simulation.
Article
Full-text available
Ambient Intelligence (AmI) is a new paradigm in information technology aimed at empowering people's capabilities by the means of digital environments that are sensitive, adaptive, and responsive to human needs, habits, gestures, and emotions. This futuristic vision of daily environment will enable innovative human-machine interactions characterized by pervasive, unobtrusive and anticipatory communications. Such innovative interaction paradigms make ambient intelligence technology a suitable candidate for developing various real life solutions, including in the health care domain. This survey will discuss the emergence of ambient intelligence (AmI) techniques in the health care domain, in order to provide the research community with the necessary background. We will examine the infrastructure and technology required for achieving the vision of ambient intelligence, such as smart environments and wearable medical devices. We will summarize of the state of the art artificial intelligence methodologies used for developing AmI system in the health care domain, including various learning techniques (for learning from user interaction), reasoning techniques (for reasoning about users' goals and intensions) and planning techniques (for planning activities and interactions). We will also discuss how AmI technology might support people affected by various physical or mental disabilities or chronic disease. Finally, we will point to some of the successful case studies in the area and we will look at the current and future challenges to draw upon the possible future research paths.
Book
Mainly deal with queueing models, but give the properties of many useful statistical distributions and algorithms for generating them.
Article
Lifestyle monitoring (LM) technology is part of a new generation of telecare which aims to observe the daily activities of older or vulnerable individuals and determine if a medical or care intervention would be beneficial. The development and validation of new LM systems should ideally involve extensive trials with users in real conditions. Unfortunately, effective user trials are very challenging, generally limited in scope and costly. In this paper, a simulator is proposed that can serve to generate synthetic data of daily activity which can then be used as a tool for the validation and development of LM systems. The most challenging part of the simulator is to replicate people's behaviour. In the paper, a novel model of daily activity simulation is proposed. Such daily activities are dependent on a number of external factors that control the need or desire to perform the activity. The proposed simulator aims to reproduce behaviour such that the probability of performing an activity increases until the need is fulfilled. It is possible to parameterise the behavioural model according to a set of features representing a particular individual. The simulator parameters have been populated using real world experiments through hardware testing and data collection with older people. Experimental verification that the desired features are reasonably reproduced by the simulator is provided.
Article
Esta obra ofrece una introducción a los conceptos, metodologías y aplicaciones de la simulación en los negocios, especialmente en el análisis de riesgos.
Article
The goals of this article are (a) to describe the daily life of the very old in terms of frequency, duration, variety, and social and physical contexts of activities, and (b) to examine the effects of background variables (e.g., age, sex, residential and marital status, income, and education) on late life activity engagement. A representative sample of 516 adults aged 70-105 was interviewed about their activities using the Yesterday Interview. In contrast to most research on activity engagement, this measurement approach allows for assessment of both the type and context of activities engaged in during the day preceding the interview. The results indicated high frequencies of obligatory activities but also showed substantial time spent in discretionary activities, with television viewing occupying most of the participants' leisure time. Most activities were done alone and at home. In bivariate and multiple regression analyses, age and residential status had the strongest association with activity frequency, duration, and variety; the oldest-old and those residing in long-term care facilities had lower levels of activity engagement. Results are discussed in terms of their relevance for successful aging.
Article
Home telemonitoring represents a patient management approach combining various information technologies for monitoring patients at distance. This study presents a systematic review of the nature and magnitude of outcomes associated with telemonitoring of four types of chronic illnesses: pulmonary conditions, diabetes, hypertension, and cardiovascular diseases. A comprehensive literature search was conducted on Medline and the Cochrane Library to identify relevant articles published between 1990 and 2006. A total of 65 empirical studies were obtained (18 pulmonary conditions, 17 diabetes, 16 cardiac diseases, 14 hypertension) mostly conducted in the United States and Europe. The magnitude and significance of the telemonitoring effects on patients' conditions (e.g., early detection of symptoms, decrease in blood pressure, adequate medication, reduced mortality) still remain inconclusive for all four chronic illnesses. However, the results of this study suggest that regardless of their nationality, socioeconomic status, or age, patients comply with telemonitoring programs and the use of technologies. Importantly, the telemonitoring effects on clinical effectiveness outcomes (e.g., decrease in the emergency visits, hospital admissions, average hospital length of stay) are more consistent in pulmonary and cardiac studies than diabetes and hypertension. Lastly, economic viability of telemonitoring was observed in very few studies and, in most cases, no in-depth cost-minimization analyses were performed. Home telemonitoring of chronic diseases seems to be a promising patient management approach that produces accurate and reliable data, empowers patients, influences their attitudes and behaviors, and potentially improves their medical conditions. Future studies need to build evidence related to its clinical effects, cost effectiveness, impacts on services utilization, and acceptance by health care providers.