Signal Processing - March 2016 - 37

and machine-learning algorithms tailored for each specific
systems implementing various tasks, such as fall detection
sensor need to be developed.
[3]-[5], mobile emergency response [6], video surveillance
[7], automation [8], monitoring activities of daily living [9],
and respiratory monitoring [10]. Falls among the elderly are a
Pir sensor signal processing
major concern for both families and medical professionals.
PIR sensors are low-cost devices designed to detect the presFalls are considered to be the eighth leading cause of death in
ence of moving bodies from stationary objects. They are easy
the United States [11] and fall injuries can result in serious
to use and can even work in the dark, unlike ordinary visioncomplications [12], [13]. Autonomous fall detection systems
based systems, because they image infrared light. A PIR senfor AL can reduce the severity of falls by informing other
sor functions by measuring the difference in infrared
people to deliver help and reduce the amount of time people
radiation between the two pyro-electric elements inside of it.
remain on the floor. These systems can increase safety and
This difference occurs due to the motion of bodies in the
independence of the elderly.
viewing range of the sensor. When the two pyro-electric eleTo truly assist elderly people, an AL system should satisfy
ments are subject to the same infrared radiation level, they
some basic requirements [14]:
generate a zero-output signal by canceling each other out.
Therefore, the analog circuitry of a PIR sensor can reject false
■ Low-cost: Almost 90% of the older adults prefer to stay in the
detections very accurately.
comfort of their own homes. Therefore, an AL system should
PIR sensors are widely used in the context of AL. In
be affordable by the average elderly person or couple.
[38],
eight PIR sensors are installed in the ceiling of hospi■ High accuracy: Since the aim is to enhance the wellness
tal rooms to assess the daily activities of
and the life quality of elderly people, a
elderly patients. The activities are classitolerable error rate should be achieved.
The
most
important
signal
fied in 24 different categories by check■ User acceptance: The AL systems
processing problem in AL
ing the number of sensors activated and
should be compatible with the ordinary
systems is the recognition recording the time interval for which they
activities of people so that they can interremain activated. Barger et al. [24] introact with the system easily, i.e., by speakof human activity from
duce a system of distributed PIR sensors
ing naturally, using simple gestures, etc.
signals generated by
Also, users do not find wearable systems
various sensors including to monitor a person's in-home activity.
The activity level of the person is defined
or those that need to be carried practical.
vibration sensors, PIR
as the number of sensor firings in a room
Thus, contact-free and remotely controlsensors, and wearable
per time spent in the room. Mixture modlable systems are desired.
accelerometers.
els are applied to the sensor data in the
■ Privacy: The AL systems should be nontraining set to develop a probabilistic
visual and share minimal private data
model of event types. These models are then used to identiwith the monitoring call center regarding the daily living
fy the type of event associated with each observation in the
activities of individuals.
test set. In [27], a PIR sensor installed in a corner of a livDespite the presence of surveys [2], [15], [16] and prolifing room is employed to detect the abnormalities in daily
eration of different types of sensors in the AL field, a comactivities of an elderly person. The PIR sensor sends the
prehensive study concentrating on the utilized sensor signal
value "1" to the controller if there are activities from the
processing methods is not available. This article aims to
person and the value "0" otherwise. Hidden Markov modprovide an overview of most recent research trends in the
els (MMs), forward algorithms, and Viterbi algorithms are
AL field by focusing on PIR sensors, vibration sensors,
used to analyze the obtained data sequence. If a certain
accelerometers, cameras, depth sensors, and microphones
deviation from the constituted models is detected, the careand the related signal processing methods, which together
giver receives an alert. In [26] a wireless sensor network
meet most of the aforementioned requirements. Ambient
including PIR, chair, bed, toilet, and couch sensors is suginformation monitoring sensors are used in home safety
gested to determine the wellness of the elderly. Time[17]-[19], home automation [8], [20]-[23], activity monitorstamped sensor activities are recorded and fed to
ing [14], [24]-[27], fall detection [28]-[34], localization and
predefined wellness functions.
tracking [35]-[37], and monitoring the health status indicaIn [25], PIR and contact sensors are used to assess neurotors of elderly and chronically diseased people outside hoslogic function in cognitively impaired elders. The contact senpitals [38]-[44].
sor is responsible for tracking the presence or the absence of
the resident and recording the time spent in the home and out
Human activity recognition
of the home. PIR sensors are utilized for the estimation of
using various sensor modalities
walking speed and daily activity. The walking speed of the
The most important signal processing problem in AL systems
resident is estimated from the time of PIR sensor firings that
is the recognition of human activity from signals generated by
are placed sequentially along a hall. The amount of daily
various sensors including vibration sensors, PIR sensors, and
activity is decided based on the number of sensor firings per
wearable accelerometers. Obviously, each sensor generates difminute when the subject is in the home.
ferent kinds of time-series data. Therefore, signal-processing
IEEE SIgnal ProcESSIng MagazInE

March 2016

Table of Contents for the Digital Edition of Signal Processing - March 2016