A CYBER-PHYSICAL SYSTEM FOR ELDERS MONITORING
Xiang Li, Ying Qiao and Hongan Wang
Intelligence Engineering Laboratory, Institute of Software Chinese Academy of Sciences
No. 4, Zhong Guan Cun South Fourth Street, Hai Dian District, Beijing, China
Keywords: Event-Condition-action Rules, Active Database, Cyber-physical System, Reasoning System, Real-time.
Abstract: In a new life style, elders can stay in multiple places instead of a single place. The new life style has two
features, i.e. multiple scenarios and changes of status of elders, which lead challenges to traditional elder
monitoring systems in cooperation and flexibility. This paper presents a new cyber-physical system for
elders monitoring, which is divided to two layers, a sub-system layer and a global service layer. Active
databases with real-time event-condition-action rule reasoning system are used as core components in sub-
systems and the global service to detect risks and cooperate with other systems actively, intelligently, and at
real-time. Structure of an ECA rule reasoning system is flexible, in which ECA rules can be adjusted on-the-
fly. We discuss some properties in the cyber-physical system, including cooperation among sub-systems,
flexibility and real-time property. At last, we present a case study to validate our works.
1 INTRODUCTION
As a person grows older, it is more and more
dangerous that he/she lives independently in the
home. Thus, many elders decide to relocate from
their homes to nursing centers. It is common,
following such a relocation, for a person to become
depressed due to their lack of independence
(Augusto and Nugent, 2004).
Over the past decade, a new life style becomes
widespread: elders stay in multiple places, e.g.
homes, nursing centers, parks, pathways, etc. instead
of a single place. Elders can be taken care by nurses
in the nursing center in the daytime and keep their
independences, privacies and personal interests in
some other time.
The life style has two features, i.e. 1) multiple
scenarios and 2) changes in the status of an elder, e.g.
healthy, indisposition, ill seriously, etc.
The two features lead two challenges:
cooperation and flexibility. Suppose an elder with
heart disease has a cold in the home. Both the
monitoring system in the home (home system) and
the monitoring system in nursing center (nursing
system) should enhance the monitoring level to
prevent a possible heart attack caused by the cold.
Here, information about the elder should be
provided by the home system to the nursig system,
so cooperation between the two systems is necessary.
Meanswhile, both systems should support to modify
contents of monitoring to fit changes of status of the
elder, so systems should be flexible.
However, traditional elder monitoring systems
are stable and independent, so they lack cooperation
mechanism and flexibility.
In this paper, we present a cyber-physical system
for elders monitoring. The system is divided to two
functional layers, sub-systems and a global service.
A sub-system is arranged in each scenario to collect
data from elders and environment, and detect risks
intelligently. The global service manage all sub-
systems. Active database with real-time event-
condition-action (ECA) rule reasoning system is
used as a core component to provide abilities of
intelligent and real-time reasoning to sub-systems
and the global service.
Through the global service, sub-systems can
cooperate with others. Meanwhile, the structure of
each active database is flexible, in which ECA rules
can be added, modified or removed on-the-fly.
The rest of the paper is organized as follows:
section 2 expresses related works; section 3
addresses the functional framework of our system;
section 4 introduces design of our system; we
validate our works in section 5; conclusions and
future works are stated in section 6.
294
Li X., Qiao Y. and Wang H. (2010).
A CYBER-PHYSICAL SYSTEM FOR ELDERS MONITORING.
In Proceedings of the 12th International Conference on Enterprise Information Systems - Databases and Information Systems Integration, pages
294-299
DOI: 10.5220/0002968402940299
Copyright
c
 SciTePress

2 RELATED WORK
Classical methods and systems in traditional elder
monitoring systems fall into three categories, i.e.,
intelligent and wearable devices, video data
processing and intelligent reasoning.
In the first category, (Park, Won, etc., 2003)
introduces several intelligent devices. (Korhonen,
Parkka, etc., 2003) presents users’ requirements and
the state of the art on wearable sensors.
In the second category, (Tseng, Wang, etc.,
2007) presents an intelligent sensor with a mobile
video camera. (Tabar, Keshavarz, etc., 2006) fuses
data from sensors and images from cameras to
monitor falling of elder.
In the third category, (Augusto and Nugent,
2004) present a set of ECA rules to monitor health
of elders. (Li, Lin, etc., 2004) presents the real-time
event detection service to handle event and data
from distributed sensor networks. (Noury, Virone,
and Creuzet, 2002) presents a decision algorithm
with rules to localize a human in the home.
Although these methods and systems have
researched on availability, reliability and efficiency
in elder monitoring, none of them consider
flexibility and intelligent cooperation among more
than one monitoring systems.
3 FUNCTIONAL FRAMEWORK
Figure 1 illustrates the functional framework of our
cyber-physical system for elders monitoring. In the
framework, the system is divided to two functional
layers, sub-systems and a global service center.
Figure 1: Function layers the cyber-physical system for
elders monitoring.
A sub-system is placed in each scenario, in order
to 1) collect data from elders and environment, 2)
detect risks from data and giving alarms actively and
at real-time, 3) exchange information with the global
service, and 4) adjust its contents of monitoring.
The global service is used to 1) receive and
synthesize data from sub-systems, 2) detect risks
globally, 3) transmit data, and 4) adjust contents of
monitoring in sub-systems.
With these functions, the cyber-physical system
will fit for the new life style. Sub-systems can
cooperate with each other by transmitting
information through the global service. Meanwhile,
monitoring contents in all sub-systems can be
adjusted by the global service.
4 DESIGN
Figure 2: Structures of the global service and sub-systems.
In order to realize aforementioned functions, sub-
systems and the global service should have abilities
of active, intelligent and real-time reasoning.
Moreover, the structure of reasoning should be
flexible so that contents of reasoning should be
adjusted. In order to support cooperation among sub-
systems and the global service, a communication
mechanism should be set up among them.
Figure 2 shows structures of sub-systems and the
global service.
In each sub-system, several kinds of sensors are
used to monitor all useful data from elders and
environment. Data received from sensors and the
global service is sent to the active database.
With ECA rules, an active database is able to
anticipate potential or actual hazardous situations
and intelligently discern how to best advise related
persons locally or what to send to the global service.
In an active database, the event filter and the
condition filter extract external events and condition
values from original data. An ECA rule reasoning
system with a real-time reasoning algorithm is used
to detect composite events, evaluate conditions, and
trigger actions at real-time depending on ECA rules.
Active database has a flexible rule base, in which
ECA rules can be adjusted on-the-fly.
A CYBER-PHYSICAL SYSTEM FOR ELDERS MONITORING
295

Some actions are implemented in the local action
implement system to react local risks. Other actions
are implemented in the data transfer to send suitable
information to the global service.
In the global service, an isomorphic active
database processes information from sub-systems
and gives out global actions. Some actions are
implemented in the global action implement system
to react for global risks. Some are implemented in
the data transfer to send suitable data to suitable sub-
system. Others are implemented in the ECA rules
base for sub-systems to choose a suitable set of ECA
rules and send them to a sub-system.
4.1 Cooperation
Information exchanged among sub-systems and the
global service is packed in XECA, a XML-based
language that we defined, including ECA rules,
events, conditions, actions, commands, etc. The
model and visual specification tools of the XECA
can be found in our previous work (Qiao, Zhong,
etc., 2007) and (Liu, Qiao, etc., 2008). Technically,
textual data can be easily passed from one site to
another, with traditional distributed technologies
(CORBA, JMI, etc.).
In the architecture, a sub-system can cooperate
with any other sub-systems indirectly through the
global service. Cooperation among systems is
transformed to cooperation among reasoning
systems. Inference results in a sub-system are
considered as external events for other reasoning
systems.
4.2 ECA Rule Reasoning System
As the core of active database, ECA rule reasoning
system can detect risks and trigger appropriate
actions actively, intelligently and at real-time. The
structure is shown in Figure 3.
In our ECA rule reasoning system, there are five
layers:
Input Interface: Through input interface, XECA-
based commands are sent into the ECA rule
reasoning system. The textual processor checks all
commands, compiles them to binary information and
sends them to correct places.
Unit-Mail Management Layer (UMML): In the layer,
commands for adjusting ECA rules are implemented
without stopping the inference process. New ECA
rules are translated to temporary units and stored in
the unit producer. Commands for adding are
implemented in the merging module by merging
temporary units into the rule base. Furthermore,
commands for removing are implemented in the
removing module, by removing invalid units in the
rule base.
Open Rule Base: The rule base holds ECA rules and
provides suitable ECA rules to the inference layer.
In the rule base, ECA rules are stored as a loose
coupling structure called unit-mail graph (UMG), in
which each composite event or each condition in
ECA rules is packed in a unit.
In the rule base, each unit is physically
independent of others, which makes it is easy to add
or remove units in the rule base. Meanwhile,
logically, units consist of a directed graph which
ensures inference can run smoothly.
Inference Layer: In the inference layer, the inference
core detects composite events from external events,
evaluates conditions and triggers actions. With
UMG, the dynamic process of inference is
transformed to a “receive, process, and send mails”
model. Details of UMG and the “receive, process,
and send mails” model can be found in (Li, Qiao,
etc., 2009).
Output Interface: In the output interface, condition
queries and actions are translated to XECA and sent
out.
Figure 3: Structure of the ECA rule reasoning system.
ICEIS 2010 - 12th International Conference on Enterprise Information Systems
296

4.3 Flexibility
With the UMML, adding and removing ECA rules
are transformed to adding and removing units in
UMG. Because the UMG is loose coupling, it is not
necessary to stop the entire reasoning system when
units are being added or removed. Four mechanisms
ensure ECA rules can be added and removed on-the-
fly:
Loose Coupling mechanism means relations
among units are loose coupling.
Lock/Unlock mechanism means a unit cannot be
called by other threads when it is locked.
Top-Down mechanism means the merging
module and the removing module add units and
delete units in top-down order.
Roll-Back mechanism makes the rule processing
mechanism avoid unexpected errors.
4.4 Real-time Algorithm
In the inference layer, an “any time” real-time
algorithm called RTIAE is used to ensure actions be
given out at real-time.
Logically, in the inference core, units consist of a
directed rule graph. In RTIAE, the reasoning is
accomplished via heuristic search. The purpose of
heuristic search is to find a path from a specific
entrance node (without incoming edges) to an exit
node (without outgoing edges) so that the time
consumed for traveling along this path is as short as
possible. During the search, the expected path will
be expanded with a node selected via the value of
the heuristic function.
The metric to evaluate the performance of the
inference algorithm is reasoning success ratio, which
is defined as the ratio of number of actions found
within a given deadline. The simulation results
demonstrate that the RTIAE takes advantages over
depth-first algorithm in term of reasoning success
ratio for variation of several parameters, i.e., the
laxity and the number of events occurring in the
system. Details of the real-time algorithm can be
found in our previous work (Qiao, Li, etc., 2008).
5 CASE STUDY
In this section, we design three cases to validate our
works.
5.1 Devices and Programs
We use some wireless sensors which are
manufactured by Crossbow Technology® to collect
data. Sensors communicate with a computer with a
gateway MIB520.
We develop the active data base in Java with
Eclipse platform. Communication mechanism is
developed with JMI. Three active databases run in
three laptops to perform as sub-systems in the home,
the nursing center and the global service respectively.
Laptops are connected by an intranet.
Figure 4: GUIs of our systems and experiment scenarios.
5.2 Case I
In Case I, we validate the function of a single
reasoning system. ECA rules are listed in Table 1.
Table 1: ECA rules for Case I.
Rule 1: (home)
On "1 hour after the stove is open" and "the elder
does not leave the home"
If the fire is still on
Do loud sounds "Risk of fire!"
Table 2: Experiment 1 for Case I.
Event
enter home open stove
t
Event
10:01:23 am 10:02:45 am
Action
Risk of Fire!
t
Action
11:02:45 am
Table 3: Experiment 2 for Case I.
Event
open stove close stove
t
Event
11:05:45 am 11:30:22 am
Action
t
Action
Results for Case I are shown in Table 2 and
Table 3. We enter home, open the stove and waits
for 1 hour. An alarm is given out expectably. Then,
we open the stove again, and close it in 1 hour. No
alarms are given out. Results show that our ECA
rule reasoning system works well.
A CYBER-PHYSICAL SYSTEM FOR ELDERS MONITORING
297

5.3 Case II
Table 4: ECA rules for Case II.
Rule 2: (home)
On “1 hour after the stove is open” and “the elder
leaves the home” and “the elder does not enter
the home”
If
the fire is still on
Do
provide data the global service “Alarm: Risk of
Fire: Home”
Rule 3: (global service)
On
“Alarm: Risk of Fire: Home” is received,
between arrival at and departure from the
nursing center
If
[null]
Do
transmit data to the nursing center “Alarm:
Risk of Fire: Home”
Rule 4: (nursing center)
On
“Alarm: Risk of Fire: Home” is received,
between arrival to and departure from the
nursing center
If
[null]
Do
loud sounds “Alarm: Risk of Fire: Home”
Rule 5: (nursing center)
On
the elder arrives the nursing center
If
[null]
Do
provide data to the global service “Data:
Arrival of the elder: Nursing Center”
Rule 6: (nursing center)
On
the elder leaves the nursing center
If
[null]
Do
provide data to the global service “Data:
Departure of the elder: Nursing Center”
In Case II, we validate cooperation among sub-
systems and the global service. ECA rules are listed
in Table 4.
Results for Case II are shown in Table 5. We
open the stove and leave home. When we arrive at
the nursing center, the sub-system notice the global
service this arrival. In 1 hour after we open the stove,
an alarm is sent by the home system to the global
service and transmitted immediately. When the
alarm is received by the nursing system, it alarms
with loud sounds. Because laptops are in the same
intranet, time for communications can be ignored.
Results show that our cooperation mechanism works
well.
5.4 Case III
In Case III, we validate ability of the global service
to adjust ECA rules in sub-systems. ECA rules are
listed in Table 6.
Table 5: Experiment 3 for Case II.
Sub-system in Home (Home)
Event
open stove leave Home
t
Event
12:11:23 pm 12:12:33 pm
Action
Home Alarm
to Global
t
Action
13:11:23 pm
The Global Service (Global)
Event
Arrival at
NC
Home Alarm
t
Event
12:30:22 pm 13:11:23 pm
Action
Alarm to NC
t
Action
13:11:23 pm
Sub-system in Nursing Center (NC)
Event
arrival at
NC
Alarm
from
Global
t
Event
12:30:22
pm
13:11:23
pm
Action
Arrival
at NC to
Global
Notify
the elder
t
Action
12:30:22
pm
13:11:23
pm
Table 6: ECA rules for Case III.
Rule 7: (nursing center)
On the elder falls over
If
[null]
Do
provide data to the global service “Data: Fall
Over: Nursing Center”
Rule 8: (global service)
On
“Data: Fall Over: Nursing Center” is received
If
[null]
Do
send two commands to home:
A command for removing Rule 9 in XECA
A command for adding Rule 10 in XECA
Rule 9: (home)
On
1 hour after the elder enter the bath room
If
[null]
Do Notify the global service “Risk of bath”
Rule 10: (home)
On
30 minutes after the elder enter the bath room
If
[null]
Do Notify the global service “Risk of bath”
Results for Case III are shown in Table 7. We
fall over in the nursing center. Then the report is sent
to the global service. The global service sends out
two commands to the home system. Then, the home
system replaces Rule 9 with Rule 10. Results show
that the function of adjusting ECA rules works well.
ICEIS 2010 - 12th International Conference on Enterprise Information Systems
298

Table 7: Experiment 4 for Case III.
Sub-system in Nursing Center (NC)
Event
falls over
t
Event
14:09:01 pm
Action
Fall over to
Global
t
Action
14:09:01 pm
The Global Service (Global)
Event
Fall over
t
Event
14:09:01
pm
Action
Command
for removing
Rule 9
Command
for adding
Rule 10
t
Action
14:09:01 pm 14:09:01 pm
Sub-system in Home (Home)
Timestamp
before
14:09:01
pm
14:09:01 pm
after
14:09:01 pm
ECA Rules
Rule 9 null Rule 10
6 CONCLUSIONS AND FUTURE
WORK
Although many works have been done on elder
monitoring systems, none of them can fit multi-
scenarios and changes of status of elders. To solve
above problems, we present a cyber-physical system
for elders monitoring. In the system, sub-systems are
used to monitor status of elders and environment,
and a global service is used to manage all sub-
systems. Active databases with real-time event-
condition-action (ECA) rule reasoning systems
provide abilities of intelligent and real-time
reasoning to sub-systems and the global service. In
comparison to traditional independent elder
monitoring systems, our cyber-physical system
enhances flexibility, cooperation among sub-systems
and real-time property. At last, we design three cases
to validate our works.
In the future, more researches could be done with
the cyber-physical system for elders monitoring.
Firstly, in our system, inferences are real-time,
but real-time property in communications among
sub-systems should be researched in the future.
Secondly, ability of description of the model of
ECA rule should be enhanced.
ACKNOWLEDGEMENTS
This work is supported by National Nature Science
Foundation of China (Grant No. 60873073).
REFERENCES
Noury, N. Virone, G., and Creuzet, T., 2002. The health
integrated smart home information system (HIS2):
Rules based system for the localization of a human. In
2nd Annual International IEEE-EMB Special Topic
Conference on Microtechnologies in Medicine &
Biology. pp. 318–321.
Park, S.H., Won, S.H., Lee, J.B., and Kim, S.W., 2003.
Smart Home: digitally engineered domestic life. In
Proceedings of the 1st international conference on
appliance design. pp. 12-17.
Korhonen, I., Parkka, J., and Gils, M., 2003. Health
monitoring in the home of the future. In IEEE
Engineering in Medicine and Biology Magazine. vol.
22, issue 3, pp 66–73. IEEE Press.
Augusto, J. and Nugent, C., 2004. A New Architecture for
Smart Homes Based on ADB and Temporal
Reasoning. In Proceedings of 2nd International
Conference on Smart Homes and Health Telematic. pp.
106-113. IOS Press.
Li, S., Lin, Y., Son, S., Stankovic, J., and Wei, Y., 2004.
Event detection services using data service
middleware in distributed sensor networks. In
Telecommunication Systems. vol. 26, No. 2-4, pp. 351-
368. Springer Press.
Tabar, A.M., Keshavarz, A., and Aghajan, H., 2006. Smart
home care network using sensor fusion and distributed
vision-based reasoning. In Proceedings of ACM
Multimedia Workshop on VSSN. pp. 145-154. ACM
Press.
Qiao, Y., Zhong, K., Wang, H., and Li, X., 2007.
Developing Event-condition-action Rules in Real-time
Active Database. In Proceedings of ACM symposium
on applied computing. pp. 511-516. ACM Press.
Tseng, Y.C., Wang, Y.C., Cheng, K.Y., and Hsieh, Y.Y.,
2007. Imouse: an integrated mobile surveillance and
wireless sensor system. In Computer. vol. 40, Issue 6,
pp. 60–66. IEEE Press.
Liu, W., Qiao, Y., Li, X., Zhong, K., Wang, H., and Dai,
G., 2008. A Visual Specification Tool for Event-
Condition-Action Rules Supporting Web-Based
Distributed System. In Proceedings of the Tenth
International Conference on Enterprise Information
Systems. pp. 246-251. Springer Press.
Qiao, Y., Li, X., Wang, H., and Zhong, K., 2008. Real-
Time Reasoning Based on Event-Condition-Action
Rules. In Proceedings: Posters of the OTM 2008
Cooperative Information Systems International
Conference. pp. 1-2. Springer Press.
Li, X., Qiao, Y., and Wang, H., 2009. A Flexible Event-
Condition-Action (ECA) Rule Processing Mechanism
based on a Dynamically Reconfigurable Structure. In
Proceedings of the Tenth International Conference on
Enterprise Information Systems. pp. 291-294. Springer
Press.
A CYBER-PHYSICAL SYSTEM FOR ELDERS MONITORING
299