Authors: Che-Chang Yang, Yeh-Liang Hsu (2007-03-16)；recommended:
Yeh-Liang Hsu (2007-08-06).
This paper is presented at the 3rd IASTED
International Conference on Telehealth: Telehealth 2007, May. 30-Jun. 1, 2007, Montréal, Canada
a Wearable System for Real-time Physical Activity Monitoring in a Home Environment
Quantitative assessment of daily physical
activity in a home environment provides significant information in evaluation
of health status and the quality of life of subjects with limited mobility and
chronic diseases. This study developed a home telehealth-based application, a
wearable system for real-time human physical activity ambulatory monitoring by using
only one mobile sensing device which utilizes tri-axial accelerometry measurement
and the distributed data processing modality. This system is able to identify
several targeted human postures, postural transitions and walking with the
embedded algorithm. In addition, this system also features fall detection
capability which might be highly desired for elderly care. The results of the
test for evaluating the performance of the system show high identification
accuracy for both still postures and dynamic movements. A long-term ambulatory
test at home was also demonstrated and the recorded data indicated sufficient
information regarding the subject’s activities of daily living. Some inherent
limitations concerning real-time identification were discussed. Despite the
limitations, this system is technically viable for ambulatory application to
provide sufficient information in evaluating a person’s activity of daily
living (ADL) and his physical mobility level. Potential wok of this system in
the future is also discussed.
Key words: Physical activity, fall detection,
tri-axial accelerometer, home telehealth, ambulatory monitoring
can be regarded as any bodily movement or posture that is produced by skeletal
muscles and results in energy expenditure [Capspersen et al., 1985].
Various health conditions such as heart disease, senile dementia, degeneration
in mobility will directly affect one’s physical activity level. However, a
majority of assessment methods used in the past mostly relied on questionnaire
registration and observation, which may possibly lead to insufficient
information and inconsistent assessment results [Wada, et al., 2005].
Therefore, quantitative assessment of daily physical activity at home is a
determinant in the evaluation of health and the quality of life of subjects
with limited mobility and chronic diseases, such as elderly persons [Foster et
assessment is difficult due to the subtle and complex nature of body movement
which requires precise and reliable monitoring approaches. Current developed
physical activity monitoring systems can be technically classified according to
the adopted sensing techniques. Ambient sensor arrays, such as various kinds of
switches, have been widely used for ADL (activity of daily living) acquisition
[Noury et al., 2002, 2003]. Such system architecture can provide
continuous monitoring in non-intrusive way. However, the inability in accurate
dynamic motion analysis has long been the major drawback that can not be
completely overcome. Although video cameras can be employed to enhance the
accuracy, such method may lead to the demanding issue of personal privacy in
The other similar
sensing technique is the use of human motion capturing systems which have been
widely used in the applications of computer animation and virtual reality.
These systems are based on optical, magnetic and ultrasonic operation
principles and can allow a complete kinematic analysis but require a dedicated
laboratory site [Aminian et al., 2004], [Bodenheimer et al.,
1997], [Disckstein, et al., 1996], [Aminian et al., 2002], [Kemp et
al., 1998]. However, complex system installation and higher cost of system
facility are the major factors unacceptable for common use in home environment.
Moreover, the subjects under monitoring must be restrained inside a
laboratory-like space, which is entirely different from a free-living home
The related research
in recent years has been focusing on developing wearable systems, which has
been regarded as appropriate alternative for physical activity monitoring [Bouten
et al, 1997], [Meijier et al., 1991], [Najafi et al.,
2002], [Veltink et al., 1996], [Karantonis et al., 2006]. In a
wearable system, the sensor units (e.g., gyroscopes or accelerometers) and
required components are integrated into portable devices, or are even
incorporated into clothing. With advancing technologies in microcontrollers and
wireless communication, real-time identification for physical activity was
achieved onboard the wearable system, without the need of off-line data
analysis in external computation phase [Karantonis et al., 2006].
The development of
home telehealth system is an emerging trend in societies of rapid aging
population [Scanaill et al., 2006]. Tele-monitoring basic health
parameters provides safe, cost-effective, efficient and patient-centered
healthcare to improve clinical outcomes at lower cost [Finkelstein et al.,
2006]. The portable tele-homecare monitoring system (PTMS) was developed by the
authors and was introduced for product-oriented design. What sets this
innovation apart from most other systems is its highly decentralized monitoring
model and the portable nature of the system. Such cost-effective and
product-oriented approach makes the system economically viable and acceptable
to the end-users [Hsu et al., 2007].
The purpose of this research is to demonstrate
a PTMS-based system for real-time monitoring and preliminary assessment of
physical activities at home. The ambulatory monitoring is achieved by
capitalizing on real-time accelerometry measurement and processing of single
wearable device. A hierarchical algorithm was developed and embedded in the
wearable device to enable real-time identification of still postures
(lying, sitting and standing), postural transitions (sit-stand, lie-sit) and walking. This system also features the
function of fall detection and its corresponding emergency alarming reports
(GSM phone message or e-mail), which may be highly desired for the elderly
care. Based-on home telehealth application, this system is implemented to
register one’s long-term physical activity data at home for preliminary
assessment of mobility level.
Methods and system design
Figure 1 shows the structure of the system. A
wearable motion detection unit (MDU) primarily consists of a miniature
tri-axial accelerometer module (KXM52-1050, Kionix,
Inc.), a PIC microcontroller (PIC18F6722,
Microchip) and a RF wireless
transmitter module (PT2262, Princeton Tech.;
TWS-CS-2, Wenshing Electronics Co., LTD).
It is designed for accelerometric measurement and real-time identification of
human postures and movements. Figure 2 shows the package of MDU and its
tri-axial orientation of acceleration measurement. The MDU uses a
gravity-responsive tri-axial accelerometer to measure acceleration and
inclination produced by human movement or posture. The sensor outputs are
initially low-pass filtered at 50Hz for noise rejection and are then
continuously sampled at 60Hz via 10-bit A/D conversion of the PIC
microcontroller. Real-time identification is simultaneously processed in the
PIC microcontroller with embedded algorithm. Each real-time identified item is
wirelessly and cyclically transmitted to a base station via RF 433.92MHz. As
shown in the Figure 2, the MDU is designed to be carried at the waist level
nearby close to the center of gravity of the body by means of clips onto the
pant belt for easier and convenient use. From the empirical trial, the suitable
position lies within the range of 45 degrees from the frontal
(antero-posterior) side to either of medio-lateral sides [Bodenheimer et al., 1997].
Figure 1. System structure
Figure 2. The attachment of MDU and the tri-axial orientations
of the sensor
The base station in the PTMS is called household
distributed data server (DDS), as shown in the Figure 3 The DDS is also a PIC
microcontroller-based device that features the capabilities of signal
computing, I/O control, Ethernet communication, wireless RF data reception,
data storage (MMC) and the links with external devices. Referring to the Figure
1, for data retrieving and management the authorized data administrators (e.g.,
doctors, care-givers, families) are allowed to access the DDS via the Internet
by using the Internet browser or application program. The DDS also features
event-driven capability for emergency alarm and report. Immediate delivery of
GSM SMS phone messages or e-mail to specific care-givers can be activated when
possible fall has been detected.
Figure 3. The household distributed data server
2.2 Algorithm design
The algorithm embedded
in the PIC microcontroller of the MDU is developed to identify three still
postures (lying still, sitting still, and standing still), four postural
transitions (sit-stand, stand-to-sit, sit-to-lie and lie-to-sit) and walking
movement. Detection for possible falls is also designed in this algorithm. Figure
4 shows the process flow of the algorithm which mainly includes five parts:
data sampling (Cx), pre-processing (Px), dynamic posture transition
identification (DBx), still posture identification (DAx)
and possible fall detection (DCx). All signals are processed in
time-domain due to the limited computation capability of the PIC
microcontroller and real-time process method. Each item is identified and
addressed in 2.5s cycle. In the case where there is no definite result
determined throughout the processes, the event will be recorded as an “Uncertain movement” or an “Uncertain posture”.
sampling process consists of the primary stage (C1) and the secondary stage (C2)
in 0.5s and 2.0s, respectively. The use of the dual-stage data sampling
strategy ensures that the data of one event can be acquired within the same
sampling interval. Initially, Sections C1
and D1 determine whether any sign of
dynamic movement exists. If no dynamic movement be detected (D1=No), the sampled 0.5s data is used
to identify one of the three possible still postures in the processes DAx.
If dynamic movement is detected (D1=Yes),
the secondary data sampling stage (C2)
is immediately activated to collect the subsequent 2.0s data. The 2.5s data
collected in both stages is combined and then median-filtered (P1, window length n=3) and simplified (P2,
one-third scaling by averaging method) to represent a “dynamic event” for the
following step-by-step identification. The “slope mapping” technique is
commonly used in most processes in this algorithm. This technique registers
whether there are apparent changes in the measured data by mapping the analog
signals into binary sequence.
Figure 4. The algorithm flowchart
In order to investigate
the accelerometric characteristics of sit-stand transitions, a test was
performed on 15 ostensibly healthy subjects in various ages arranged in three
groups: Young (20-35 yrs), middle-aged (35-50 yrs) and elderly (50+yrs), with 5
subjects in each group. From the observation of the test, the vertical
acceleration is used to identify sit-stand postural transitions in which the
acceleration patterns can be characterized by three particular rules: (i) peak
order, (ii) peak distance (time interval) and (iii) peak values. Either of
sit-stand postural transitions can only be identified when the criterion of all
the three rules are satisfied. The lie-sit postural transitions can be identified
and further distinguished from each other by investigating both the trunk tilt
variation and the final posture orientation. Still posture identification
requires the information of previously known postural transitions or walking
movement. A lying still posture can be recognized according to the posture orientation
or if there exists a previous sit-to-lie postural transition. Similarly,
sitting still or standing still postures can be identified by the existence of
the types of previous sit-stand transitions or walking movement. Walking can be
recognized in a regular oscillating form in vertical acceleration. Fall is
regarded as a “sign of fall” occurs and is followed by a prolonged lying
posture. The sum of tri-axial acceleration values is the determinant for recognizing
a “sign of fall”.
For the evaluation
of the performance of the algorithm, 10 subjects were recruited for the
laboratory-based test. Sensitivity and specificity tests for posture (lying
still) and posture transitions (sit-stand transitions, lie-sit transitions) and
walking movement were conducted. Note that the evaluation did not include the
sitting still or standing still postures due to the fact that both still
postures are associated with the results of previously identified postural
transitions or movement. In addition, Falling was not included because it was
not easy for the testers to simulate “standardized” falls. Table 1 shows the
evaluation results of sensitivity and specificity from 200 and 500 samples,
Table 1. Performance of the algorithm
The DDS is
accessible via the Internet and the real-time status and the counts of each monitored
item can be displayed on the Internet browser (e.g., IE ) of the client PC. In
addition, a VB-based application program for this system was also developed for
complete data management, as shown in the Figure 5. This application program
features the following functions:
Data access via the Internet
authorized users (e.g., the system administrator, care-giver or families) are
allowed to access to the DDS and retrieve the data stored in it through TCP
Real-time monitoring status
The real-time monitoring information (still
posture or dynamic activity) is displayed on this interface.
The long-term recorded data in a specific
period can be chronologically displayed. Rests and activities can be distinguished
and the numbers of counts, percentages of each identified item are shown. All
the related information can be saved to an Excel file (*.xls) in the client PC.
The DDS can be optionally equipped with a GSM
module to provide event-driven capability that can be enabled when a possible
fall has been detected. A text message is sent to a user-defined client for
Figure. 5 (a) The VB application program for data
management. (b) An example of continuous monitoring
A test on continuous
ambulatory monitoring at home is conducted with data registered from about 1:00
to 16:00. as shown in Figure 5(b). In the home environment, the test user is
not expected to use this system throughout the monitoring period during some
situations such as taking a shower, going outside, etc. Therefore, the data
recorded may not be constantly continuous, and the actual monitoring time in
this test is about 405 minutes, or about 6.75 hours within the span of 15-hour
period. Figure 6 is the activity chronograph which chronologically displays the
entire recorded events. Lines of different length (also assigned different
number) represent respective identified events among the nine item design. A
report was made by the subject to register his actual activities under that
monitoring period. The monitored data also reveals good correlation with
reference to the results from the report.
Figure 6. Example of activity chronograph of the
continuous monitored data
Discussion and conclusion
A physical activity monitoring system which
utilizes only one wearable sensing device has been developed and demonstrated
in ambulatory tests. An advanced tri-axial accelerometer was used in this study
to measure the acceleration and trunk tilt of the human body. In fact, the most
precise tilt sensing can be maintained when the accelerometer is at static, or
under constant acceleration. Tilt sensing using accelerometers still has
limited accuracy in changing acceleration magnitude [Elbl, 2005]. However, In
spite of this constraint, tilt sensing and acceleration measurement using one
tri-axial accelerometer is still valid for physical activity because the
resulting outputs still preserve apparent characteristics for either trunk tilt
and acceleration patterns.
The wearable MDU has been designed to measure human
body movement at waist level and clipped to the belt for minimizing discomfort
and inconvenience in use. Power consumption which has significant influence on
the effective distance and stability of data delivery is the major issue
regarding continuous monitoring. The limitation in computation capability and
memory capacity of the microcontroller used in this study, coupled with the
fact that human events must be identified simultaneously to keep up with the
next data acquisition process, limit the identification performance. Most other
off-line systems use powerful PC-based computation software such as MATLAB to
analyze the recorded data. Therefore, identification accuracy of those systems is
usually higher than that of the real-time systems , .Despite the fact
that the algorithm achieves good performance for laboratory-set tests, those
factors may be minimized with advancing progress in microcontrollers and
wireless data communication technologies.
In this study, a
wearable system for real-time physical activity monitoring using single
wearable sensing device was developed for in-home ambulatory monitoring. This
system is able to distinguish rests from activities and further identify
several target postural transitions and walking. Although the nature of actual
human postures and activities of daily living are more complex than what is
considered and assumed in the algorithm, this algorithm still exhibits
acceptable performance in determining those target postures and activities.
Despite some limitations in the configuration for real-time data processing,
this system is technically viable to perform long-term ambulatory monitoring in
a home environment and to provide sufficient information in evaluating a person’s
activities of daily living (ADLs) and his status of physical mobility. In the
future, the application field of this system, system robustness and reliability
and the possibility for ubiquitous computing which integrates all the
ADL-related data altogether should be further considered.
Aminian, K., Najafi, B., Bula, C., Leyvraz, P.
F., 2002. “Spatial-temporal Parameters of Gait Measured by an Ambulatory System
Using Miniature Gyroscope,” Journal of
Biomechanics, Vol.35, pp.689-699.
Aminian, K., Najafi, B., 2004. “Capturing Human
Motion Using Body-Fixed Sensors: Outdoor Measurement and Clinical Applications,”
Comp. Anim. Virtual Worlds, Vol.15,
Bouten, C. V., Koekkoek, K. T. M., Verduin, M.,
Kodde, R., Janssen, J. D., 1997. “A Triaxial Accelerometer and Portable Data
Processing Unit for the Assessment of Daily Physical Activity,” IEEE Trans. Biomed. Eng., Vol.44, No.3, pp.136-147.
Bodenheimer, B., Rose, C., Thalmann, D., 1997. “The
Process of Motion Capture: Dealing with the Data,” Proceeding of Eurographicsa Computer Animation and Simulation’97.
Capspersen, C. J., Powell, K. E., Christenson, G.
M., 1985. “Physical Activity, Exercise and Physical Fitness: Definitions and
Distinctions for Health-related Research,” Public
Health Rep., Vol.110, pp.126-131.
Dickstein, R., Abulaffio, N., Gelernter, I., and Pillar, T., 1996. “An Ultrasonic-operated
Kinematic Measurement System for Assessment of Stance Balance in the Clinic,” Clinical Biomechanics, Vol.11, No.3, pp.173-175.
Elble, J. R., 2005. “Gravitational Artifact in
Accelerometric Measurements of Tremour,” Clinical
Neurophysiology, Vol.116, No.7, pp. 1638-1643.
Finkelstein, S. M., Speedie, S. M., Potthoff,
S., 2006. “Home Telehealth Improves Clinical Outcomes at Lower Cost for Home Healthcare,”
Telemedicine and e-health1, Vol.12,
Foerster, F., Smeja, M., and Fahrenberg, J.,
1999. “Detection of Posture and Motion by Accelerometry: A Validation Study in
Ambulatory Monitoring,” Comput. Hum. Beh.,
Vol.15, No.5, pp.571-583.
Hsu, Y. L., Yang, C.C., 2007. “Development of a
Decentralized Telehomecare Monitoring System”, Journal of Telemedicine and e-health, Vol.13, No.1, pp.69-77.
Karantonis, D. M., Narayanan, M. R., Mathie, M.,
Lovell, N. H., Celler,B. G., 2006 “Implementation of a Real-Time Human Movement
Classifier Using a Triaxial Accelerometer for Ambulatory Monitoring,” IEEE Trans. Info. Tech Biomed., Vol.10, No.1,
Kemp, B., Janssen, A., Kamp, B. V. D., 1998. “Body
Position Can Be Monitored in 3D Using Miniature Accelerometers and Earth-magnetic
Field Sensors,” Electroencephalography
and clinical Neurophysiology, Vol.109, No.6, pp.484-488.
Meijer, G. A. L., Westerterp, K. R., Verhoeven,
F. M. H., Koper, H. B. M., and Hoor, F. T., 1991. “Methods to Assess Physical
Activity with Special Reference to Motion Sensors and Accelerometers,” IEEE Trans. On Biomed. Eng., Vol.38, No.3, pp.221-229.
Mizuno, K., Wada, T., Kita, T., Sakagami, T., Ishine,
M., Okumiya, K., Matsubayashi, K., Rambo, T. A., 2005. “Depression, Activities
of daily living, and quality of life of community-dwelling elderly in three
Asian countries: Indonesia, Vietnam, and Japan,” Archives of Gerontology and Geriatrics, Vol. 41, No.3, pp.271-280.
Najafi, B., Aminian, K., 2002. “Measurement of
Stand-Sit and Sit-Stand Transitions Using a Miniature Gyroscope and Its
Application in Fall Risk Evaluation in the Elderly,” IEEE Trans. Biomed. Eng., Vol. 49, No.8, pp.843-851.
Noury, N., Virone, G., Creuzet, T., 2002. “The
Health Integrated Smart Home Information System(HIS2): Rules Based
System for The Localization of a Human,” Proceeding
of the 2nd Annual International IEEE-EMBS Special Topic Conference
on Microtechnologies in Medicine & Biology, pp.318-321.
Noury, N., Virone, G., Barralon, P., 2003. “New
Trends in Health Smart Homes,” Proc. 5th
International Workshop on Enterprise
Networking and Computing in Healthcare Industry, pp.118-127.
Scanaill, C. N., Carew, S., Barralon, P., Noury,
N., Lyons, D., and Lyons, G. M., 2006. “A review of approaches to mobility
telemonitoring of the elderly in their living environment,” Annals of biomed. Eng. vol.34, No.4, pp.547-563.
Veltink, P. H., Bussmann, H. B. J., Vries, W.
D., Martens, W. L. J., Lummel, R. C. V., 1996. “Detection of static and dynamic
activities using uniaxial accelerometers,” IEEE
Trans. Rehab. Eng.,
Vol. 4, No.4, pp.375-385.