Smart Home Security System Using IoT

Smart Home Security System using IoT

Abdul Rohman1, Ananda Fikri N2, Andikha Wira P3, Ikhsan Galih R4,
Iwan Arif S5, Muhammad Rifqi A6, Stefanus Willy P7 , Zido Yuwazama8
1,2,3,4,5,6,7,8
Engineering Physic, Department of Nuclear Engineering and
Engineering Physics, Universitas Gadjah Mada

Abstract A smart home is a home, usually the new one that is equipped with
special structured wiring to enable occupants to remotely control or program an
array of automatic home electronic device by entering single command [1]. Smart
home technology can be seen from many things. For example in the thermal
system (air conditioner), energy management, smart home network, smart home
sensor, etc. A smart home security system offers many more benefits than the
conventional security system.
Keyword IoT, Security System, Smart Home Security

I. Introduction source depends on the deployment

A smart home offers many more
benefits. One of the benefits is helping
homeowner to do monitoring security
even when the homeowner is not in the
house. The system uses home
automation technology which is
connected to the sensors inside the
home. If something happens in the
home, the sensor will send notification
to the homeowner’s android or IOS
using IoT system, so that they can
investigate what happen in the house. In
this paper, we will discuss smart home
security system for monitoring the
house situation.
environment and the availability of
power for applications such as solar
panel based outdoor temperature
sensing. In a replaceable battery, the
sending power and sampling rate of
data decide battery life while the energy
harvesting applications use the external
energy source, so they are somewhat
free from the battery usage method.
Wireless sensors and networks are
comprised of several nodes prepared
through various sensing devices
(sensors, controllers, and actuators) and
RF chips for wireless communication
[2].

II. Smart Home Technology As the elderly people living alone

are enormously increasing recently, we
Home automation and monitoring
need the system inferring activities of
are the dominant applications of WSNs
daily living (ADL) for maintaining
(Wireless Sensor Network), where a
healthy life and recognizing emergency.
number of heterogeneous sensors are
The system should be constructed with
deployed, to determine different
sensors, which are used to associate
activities of inhabitants. Wireless
with people's living while remaining as
sensors can be operated through
non-intrusive views as possible. Not
batteries as well as plugging into the
only that, smart home also uses smart
power supply. The selection of power
1 |

home network. By connecting some
devices or some electronic stools with
our pc or android, we can have some
information easily as in Fig. 1. It has to
be programmed with some command
and then connected with technology
called internet, and we can access by
our android. Through smart home, the
device can turn on or off automatically
and can save the energy that we call
energy management as in Fig. 2. The
main reason is that the use of modern
automation technology in home
scenarios promises considerable
savings of energy, therefore,
simultaneously reducing the
operational costs of the building over its
whole lifecycle.
Fig.1 Smart Home Integration Service
Fig.2 Smart Home Device
Smart Home Security System Using IoT
III. IoT-based Monitoring System
As the technology mature and the
interest on the internet increases,
network security has become a major
concern to companies throughout the
world. The fact that the information and
tools needed to penetrate the security of
corporate networks are widely available
has increased that concern. Because of
this increased focus on network
security, network administrators often
spend more effort protecting their
networks than on actual network setup
and administration. Tools that probe for
system vulnerabilities, such as the
Security Administrator Tool for
Analyzing Networks, and some of the
newly available scanning and intrusion
detection packages and appliances,
assist in these efforts, but these tools
only point out areas of weakness and
may not provide a means to protect
networks from all possible attacks.
Thus, as a network administrator, you
must constantly try to keep abreast of
the large number of security issues
confronting you in today’s world [1].
Monitoring the smart house can be seen
using IoT technology as in Fig. 3. The
homeowners can watch their house
condition by using CCTV from their
android or IOS system.
Fig.3 Smart Home Network
2 |

IV. Smart Home Security System
Besides using IoT for enhancing the
home security system, sensors can be
applied to the security system as well.
Fig. 3 shows the architecture of a
network-based fire detection system
that is able to overcome the
shortcomings of a conventional fire-
detection system. As shown in the
figure, a fire detector, an actuator, a bell,
and a display device are connected by a
shared transmission medium, and
information is exchanged using digital
communication. Using this connection
method, the receiver can identify which
fire detector senses a fire because each
fire detector has its own address. Also,
because the receiver periodically
examines the state of the fire detectors,
it can recognize a breakdown in the
system, such as the failure of a fire
detector or an open circuit in the
transmission medium. In addition, the
number of false alarms is less than with
conventional systems because analog
data such as the quantity of smoke and
the amount of heat measured by each
fire detector can be sent to the receiver.
It is straightforward to apply such a
system to an intelligent fire detection
system and to apply a reasoning
algorithm, because the receiver can
make use of the digitized detection
values of the smoke and heat from
multiple fire detectors installed in the
same guard area Fig. 4. As changes in
the detection signals can be calculated
and compared, assessment of risk can
be improved [5].
Fig.4 Structure of a Network-based
fire-detection system
Smart Home Security System Using IoT
Besides, a PC-based receiver can be
used in a network-based fire-detection
system for greater convenience.
Generally, it is easier to install and
maintain a PC-based receiver than a
conventional dedicated receiver. Since
man-machine interface (MMI)
technology can be applied to a PC-
based receiver, it is straightforward to
create a user interface and integrate the
fire detection system into a home
network system. Recently, several
protocols, such as BACNet, LonWorks,
and Bluetooth, have been developed to
implement network-based fire detection
systems in intelligent buildings [3] [4].
V. Conclusion
Smart Home is a residence that
uses a Home Controller to integrate the
residence's various home automation
systems. Integrating the home systems
allows them to communicate with one
another through the home controller,
thereby enabling single button and
voice control of the various home
systems simultaneously, in
preprogrammed scenarios or operating
modes. Security has been an important
issue in the smart home application.
Reference
[1] Robles, Rosslin John, et al. "A
review on security in smart
home
development." International
Journal of Advanced Science
and Technology 15 (2010).
[2] Challoo, Rajab, A. Oladeinde, Nuri
Yilmazer, Selahattin Ozcelik,
and L. Challoo. "An overview
and assessment of wireless
technologies and co-existence
of ZigBee, Bluetooth and Wi-Fi
devices." Procedia Computer
Science 12 (2012): 386-391.
3 |
 Smart Home Security System Using IoT

[3] Yanfei Li, Meiqin Liu, Weihua

Sheng, "Indoor human tracking
and state estimation by fusing
environmental sensors and
wearable sensors", Cyber
Technology in Automation
Control and Intelligent Systems
(CYBER) 2015 IEEE
International Conference on, pp.
1468-1473, 2015.
[4] B. Fong, L. Situ, L. C. K. Poon, J.
Liu, R. T. Mo, K. F. Tsang, "A
prognostics framework for
reliability optimization of mass-
produced vehicle onboard
diagnostics system", Consumer
Electronics (GCCE) 2015 IEEE
4th Global Conference on, pp.
408-409, 2015.
[5] Kyung Chang Lee and Hong-Hee
Lee, "Network-based fire-
detection system via controller
area network for smart home
automation," in IEEE
Transactions on Consumer
Electronics, vol. 50, no. 4, pp.
1093-1100, Nov. 2004.

4 |
 2015 IEEE 4th Global Conference on Consumer Electronics (GCCE)

A Prognostics Framework for Reliability Optimization of Mass-
Produced Vehicle Onboard Diagnostics System
B. Fong, Senior Member, IEEE, L. Situ, L. C. K. Poon, J. Liu, R. T. Mo and K. F. Tsang, Senior
Member IEEE
The Automotive Parts and Accessory Systems R&D Centre bfong@ieee.org
Abstract—Assistive automobile and infortainment systems
have been fitted as standard equipment in most modern
medium to high end motor vehicles. While these system offer
numerous advantages to consumers in enhancing operability
and driving safety, there is little incentive for motor vehicle
manufacturers to adopt such technology unless their return-on-
investment (ROI) can be fully realized. This paper proposes a
prognostics and health management (PHM) based solution for
profit enhancement to the manufacturers as well as to improve
consumer experience.
I. INTRODUCTION
Modern motor vehicles equipped with numerous self-
diagnostic modules possess a great challenges for optimizing
reliability of controller area network (CAN) that serves as a
backbone for intra-vehicular communications. Each of these
modules may work independently with different
communication protocols [1]. These modules that operate
under very different environments serve different functions
are subjected to very different conditions that may have a
substantial impact on their remaining useful life (RUL)
estimation [2]. Any reliability assessment of vehicle
diagnostics system cannot be carried out by adopting
standard practices commonly documented in traditional
reliability handbooks since these methods are unable to take
into consideration the actual use conditions of individual
motor vehicles [3]. A prognostics approach using data
collected from sensors mounted in different parts of the
vehicle is therefore proposed to analyze the actual use
profile to more thoroughly address the optimization of
operational reliability [4]. The proposed framework is
optimized for mass-produced vehicle onboard diagnostics
systems with a focus on maximizing return-on-investment
(ROI) for vehicle service centers.
II. ONBOARD DIAGNOSTICS
An onboard diagnostics (OBD) system consists of a
sensing network as shown in Fig. 1, these sensors collect
data for infotainment, continuous vehicle health monitoring,
passive and active safety. This work develops a self-
configuring and adaptive architecture based on a prognostic
network health management model whose main purpose is to
interconnection between entities within a vehicle-to-
infrastructure (V2I) wireless network backbone, including
individual vehicles, service centers and emergency response
centers, in such a way as to make them fully-compliant with
the EU eCall statutory regulation with enhanced security and
reliability [5].
This work is supported by funding from the Hong Kong Productivity
Council.
978-1-4799-8751-1/15/$31.00 ©2015 IEEE
.
Fig. 1. An in-vehicle controller area network that consists of different
sensors for infotainment, vehicle health monitoring, passive and active
safety
III. PROGNOSTICS FOR OPTIMAL MAINTENANCE
A. Prognostics Framework
The concept of prognostic and network health
management has been adapted in the optimization of
wireless telemedicine systems for life-saving
applications making it particularly suitable for
deployment in an eCall vehicular system [6]. Such
process entails the classifications of various precursors
for ontology generation [7], which in turn provides the
necessary information for fault diagnosis [8].
B. Operation
The optimization algorithm relies on a data acquisition
event received under any one the following three conditions:
on checking sensors of the vehicle when the engine is started
and no data is received from the driver’s seat (i.e. assume
the vehicle is idle and driver is outside the vehicle); on
receiving data from the driver’s biosensors; on receiving
data from onboard sensors when the vehicle is started and
the driver’s biosensors are active. An anomaly event is the
opposite case, i.e. it occurs under the following three
conditions: on receiving no signal from biosensors when the
driver’s seat detects presence of a driver; on detecting signal
interruption from onboard sensors; on connecting to the
CAN when the OBD is already in receiving data from both
driver and vehicle. The CAN’s internal state transition
according to the occurrence of those anomaly events are
detected to generate an alert. To handle uncertainties
associated with the wireless network connection while the
vehicle is running, the OBD attempts to send a request
message over a predefined period of time until an ACK is
received from the response centre or timeout elapses. The
reliability optimization and control algorithm of the
prognostics system is listed in Section C.
408
 C. Optimization Algorithm IV. BUSINESS INCENTIVES
Health is the current health state of any predefined There are numerous business advantages for motor
precursor, such that: vehicle manufacturers especially concerning authorized
(Health can be Normal, Abnormal or Marginal) dealers’ service centres. The OBD collects failure codes
State is the internal state of the CAN,
(State can be: Ready, Send Request or Recovery Request)
from individual vehicles, and generate a detailed description
Engine ignition; of the associated fault so that the health of the vehicle can be
IF DATA enters the OBD from the driver’s seat remotely assessed. Necessary replacement parts are ordered
Health = ‘Driver’ based on detected needs such that redundant inventory can
IF State is ‘Ready’ or ‘Send Request’
be minimized [11]. The system provides enhancement
Recovery request; wake OBD, turn monitor and biosensors on
ELSE DATA enters the OBD from the vehicle sensing network features that reduces operational cost without altering
IF Health is ‘vehicle’ existing vehicle engine or any major components. It will
IF State is ‘Ready’ or ‘Recovery Request’ transmit vehicle operating and diagnostics information to the
Send request; make CAN sleep, turn driver’s monitor and sensors off
server, and automatically schedule and generate reminders
ELSE IF Health is ‘driver’
IF State is ‘Ready’ or ‘Send Request’ for servicing. The collected data can also can provide more
Recovery request; wake CAN, turn driver’s monitor and sensors on customized service solutions to their clients through business
Exit; intelligence technology and supports the development of a
IF the OBD exits the monitoring of driver’s state
systematic approach to customer relationship management
Health = ‘vehicle’
IF State is ‘Ready’ or ‘Recovery Request’ [12] to reduce loss of revenue due to vehicle owners not
Send request; make CAN sleep, turn monitor and sensors off returning vehicles to authorized service centre for servicing.
ELSE IF the OBD exits the monitoring of vehicle’s state
IF Health is ‘driver’ V. CONCLUSION
IF State is ‘Ready’ or ‘Send Request’
Recovery request; wake CAN, turn monitor and sensors on A prognostics based in-vehicle onboard diagnostics
ELSE IF Health is ‘vehicle’ system is presented utilizing intra-vehicular wireless network
IF State is ‘Ready’ or ‘Recovery Request’ with added features of driver and vehicle health monitoring
Send request; make CAN sleep, turn monitor and sensors off"
in addition to supporting elementary EU eCall service
TABLE I requirements. The OBD interconnects between entities
SENSING NETWORK PARAMETERS within a vehicle-to-infrastructure (V2I) wireless network
Network Parameter Specification backbone for vehicles, service centres and emergency
Biosensors Transmission Power 2 dBm response centres.
Driver Health Packet Size 512 Bytes
(Wearable) Packet Inter-arrival Time > 10 ms REFERENCES
Nominal BER 10-4
Onboard sensors Transmission Power 5 dBm [1] K. C. Lee and H. H. Lee, “Network-based fire-detection system via
Vehicle Health Nominal BER 10-3 controller area network for smart home automation”. IEEE
(Fixed, Fig. 2) Shared Bandwidth 1 MHz Transactions on Consumer Electronics, vol. 50, no. 4, Nov 2004, pp.
Carrier frequency 2.410 GHz (Centre) 1093-1100.
Controller Area Transmission Power 10 dBm [2] P. A. Sandborn and C. Wilkinson, “A maintenance planning and
Network (CAN) Packet Size 8 192 Bytes business case development model for the application of prognostics
Data Rate 11 Mb/s and health management to electronic systems”. Microelectronics
Network Bandwidth 4 MHz Reliability, vol. 47, no. 12, Dec 2007, pp. 1889-1901.
Center frequency 2.4 GHz [3] D. Kececioglu, “Reliability and life testing handbook (Vol. 2)”.
DEStech Publications, 2002.
[4] D. K. Lau and B. Fong, "Prognostics and Health Management,"
One of the main functions of the OBD is to control the Microelectronics Reliability, vol. 51, no. 2, Feb 2011, pp. 253-254.
state of the CAN and detects anomaly based on data [5] J. Jääskeläinen, “Cooperative Mobility and Connected Car The
European Strategy”. Car Connectivity Consortium Summit. Nov 2012.
received from the two sensing networks, namely the driver’s [6] B. Fong, N. Ansari and A. C. M. Fong, “Prognostics and health
biosensors for driver health and onboard sensors for vehicle management for wireless telemedicine networks”, IEEE Wireless
health with specifications summarized in Table I. An Communications, vol. 19, no. 5, Oct 2012, pp. 83-89.
additional feature of the OBD is to address the EU [7] Q. T. Tho, et. Al., “Automatic fuzzy ontology generation for semantic
web”. IEEE Transactions on Knowledge and Data Engineering, vol.
requirement of collision monitoring and alerting through 18, no. 6, Jun 2006, pp 842-856.
anomaly detection. Whenever the OBD state is changed, the [8] T. T. Quan, S. C. Hui and A. C. M. Fong, “Automatic fuzzy ontology
OBD reports the detected anomaly to the response centre generation for semantic help-desk support”. IEEE Transactions on
Industrial Informatics, vol. 2, no. 3, Jun 2006, pp. 155-164.
server as an event either as an emergency or impeding [9] K. Wiens, “Dozuki Brings Technical Manuals into the 21st Century:
failure for maintenance scheduling. A Breakthrough in the World of Documentation”, IEEE Consumer
Electronics Magazine, vol. 1, no. 1, Feb 2012, pp. 39-42.
D. Scheduled Maintenance [10] B. Fong, and C. K. Li, “Methods for Assessing Product Reliability:
Upon detection of a non-emergency abnormal event, Looking for enhancements by adopting condition-based monitoring”.
IEEE Consumer Electronics Magazine, vol. 1, no. 1, 2012, pp 43-48.
maintenance work can be scheduled with necessary [11] S. C. Hui, A. C. M. Fong, and G. Jha, “A web-based intelligent fault
replacement parts ordered [9]. Diagnostics of individual diagnosis system for customer service support”, Engineering
modules are carried out using standard reliability assessment Applications of Artificial Intelligence, vol. 14, no. 4, Aug 2001, pp.
537-548.
methods with a condition-based monitoring approach [10]. [12] K. P. Jantke, J. Krebs and M. Santoso, “Game amusement & CRM:
This allows optimal scheduling to minimize downtime. Castle Scharfenstein AR case study”, ICCE GCCE 2014, pp. 488-491.

409
 Available online at www.sciencedirect.com

Procedia Computer Science 12 (2012) 386 – 391

Complex Adaptive Systems, Publication 2

Cihan H. Dagli, Editor in Chief
Conference Organized by Missouri University of Science and Technology
2012- Washington D.C.

An Overview and Assessment of Wireless Technologies and Co-

existence of ZigBee, Bluetooth and Wi-Fi Devices
R. Challooa, A. Oladeindea, N. Yilmazera, S. Ozcelikb, L.Challoo c
a
Department of Electrical Engineering and Computer Science
b
Department of Mechanical and Industrial Engineering
c
College of Graduate Studies and Department of Educational Leadership & Counseling
Texas A&M University-Kingsville, 700 University Blvd., Kingsville, TX 78363, USA

Abstract

Wi-Fi, ZigBee and Bluetooth wireless communication systems utilize the Industrial Scientific and Medical-(ISM) Band, which
results in a high mutual interference between these technologies since they all these systems operate at the same or very close
frequency bands. The interference problem increases with an in-device Co-existence (technologies existing on same device). This
is primarily due to the characteristics of each technology such as access mechanism, frame structure, peak transmit power and
frequency of operation. This work describes the interference between the Wi-Fi mostly as an aggressor on Bluetooth and ZigBee
wireless networks. So the experimental analysis of the coexistence of these three technologies in an assumed home environment
is studied especially when ZigBee is enabled for a Home Automation Network where there could be close proximity of Wi-Fi
and Bluetooth devices such as PDAs and mobile phones. The obtained result shows that there is severe degradation on ZigBee
and Bluetooth packet transmission of packets as well as re-transmission of ZigBee packets when Wi-Fi is operating.

Keywords: Wi-Fi, ZigBee, Bluetooth, ISM, aggressor, Home Automation Network, PDAs

1. Introduction

ZigBee (IEEE 802.15.4) is establishing an enabling place for the Wireless Sensor Network (WSN) especially in
the application of home automation network because of its low power and cost. Its lower power is vulnerable to
other wireless technology that has higher power and working in same spectrum. It’s among the wireless
technologies that share the unlicensed ISM band used purposely for Personal WANs. However, the Bluetooth (IEEE
802.15.1) and Wi-Fi (IEEE 802.11) share the same unlicensed band with the ZigBee and experience mutual
interference problem especially at close proximity. Adaptive Frequency Hopping, a spread-spectrum technique, was
introduced to Bluetooth technology to mitigate the problem of interference between it and Wi-Fi and any other
technology that exist in same spectrum [1]. According to [2], [3] IEEE 802.15.4 has a little impact on the IEEE
802.11 performance. However, IEEE 802.11 can have a serious impact on the ZigBee and Bluetooth performance if

1877-0509 © 2012 Published by Elsevier B.V. Selection and/or peer-review under responsibility of Missouri University of Science and Technology.
Open access under CC BY-NC-ND license. doi:10.1016/j.procs.2012.09.091
 R. Challoo et al. / Procedia Computer Science 12 (2012) 386 – 391 387

the channels allocation is not carefully taken into account. Mutual interference among these technologies is
considered a topical issue especially among technologies that shares spectrum that are close enough to cause
interference (< 50MHz channel separation). Research has been done to find the best practice to alleviate the
interference among devices using technology in same spectrum. This paper describe existing work and experimental
results on the IEEE 802.11 Wi-Fi standards, its interference with the ZigBee and the Bluetooth technologies in a
home environment where ZigBee is enabled for a home automation while Wi-Fi and Bluetooth devises such as
PDAs and mobile phones, gaming devices co-exist with it in same home.

2. Technology Overview

A quick overview of the three technologies was discussed. Aspects of the three technologies that are necessary
for the full comprehension of this study were also discussed such as the number of channels, the transmission power,
modulation type and the access scheme.

2.1 Wi-Fi
The IEEE 802.11 network is a specification of the Wireless Local Area Network (WLAN). In its low band mode,
IEEE 802.11(b, g, n) transmit data from 11 Mbps and up to 54M bps and goes up to 32 meters indoors and 95
meters outdoor [11]. The IEEE 802.11n standard uses double the radio spectrum compared to 802.11a or 802.11g.
However, IEEE 802.11a, c transmit data is up to Gbps and can exceed range by more than two times of the b and g
technologies. Low band Wi-Fi transmits in the ISM 2.4 GHz band while the high band transmit in the 5 GHz band.
A BPSK and QPSK digital modulation technique is used to transmit data up to 54 Mbps and each channel in the
ISM band is 22 MHz wide and are overlapped. Any two channels whose channel numbers differ by five or more do
not overlap. Wi-Fi’s Enhanced Isotropic Radiated Power (EIRP) is limited to 20 dBm (100 mW) [11].

2.2 Bluetooth
The IEEE 802.15.1 standard is a proprietary open wireless technology standard for exchanging data over a short
distance. It uses the short wavelength radio transmission ISM Band in the 2400-2480 MHz. It is desired for the
Wireless Personal Area Network WPAN adopted solely to replace the cable technology. Bluetooth radio adopts the
Frequency Hopping Spread Spectrum – FHSS. It occupies the entire ISM band thereby utilizing 79 channels with
each channel at 1 MHz. A GFSK, EDR, ʌ/4-DQPSK and 8 DPSK modulation formats are employed in the
Bluetooth Technology. The transmit distance of this technology ranges with the transmit power. Class 1 device of
output power of 100 mW transmit up to 100 meters while device of 25 mW output power transmit can reach up to
10 meters [11].

2.3 ZigBee
This specification was adopted for a low cost, low power digital radios and had found application in areas like
home automation, telecommunication services, healthcare and remote control just to mention a few. Similar to the
Wi-Fi and the Bluetooth technologies, ZigBee also operates in the ISM radio band. Data transmission rates of 250
Kbps [6]. ZigBee (IEEE 802.15.4) technology specifies the physical and medium access control layers for low rate
wireless PANs and transmits up to 10 meters [7]. Sixteen channels are defined for this specification in the 2.4 GHz
band but with a narrower band of 2 MHz and also do not overlap. So, up to sixteen ZigBee network can coexist in
same area and at the same time. A latest ZigBee release supports frequency hopping in the “ZigBee Pro” Standard.
This allows a ZigBee PAN to move from one channel to the other if overloading occurs in the former channel [6].
The communication model requires it to distribute work among many different devices which resides within
individual ZigBee nodes which in turn forms a network.

2.4 Channel, Frequency and Modulations

ZigBee channel is 2 MHz wide, as opposed to a Wi-Fi channel which has bandwidth of 22 MHz.. The Bluetooth
channel is 1MHz wide and occupies the whole available spectrum and possess high frequency hopping rate. Figure 1
shows the allocation of the ZigBee and Wi-Fi channels over the 2.4 GHz ISM band. The Wi-Fi individual channel
overlaps four ZigBee channels while the three most used non-overlapping Wi-Fi channels are channels 1, 6 and 11
which leaves channels 25 and 26 of the ZigBee channels free of interference. Also, the maximum allowable transmit
power of the Wi-Fi output which could be up to 100 times higher than the maximum allowed power of the ZigBee
388 R. Challoo et al. / Procedia Computer Science 12 (2012) 386 – 391

could also serve as a further aspect making the coexistence of Wi-Fi and ZigBee difficult.

Figure 1: Allocation of ZigBee and Wi-Fi Channels over the ISM Band [11]
3. BLUETOOTH AND Wi-Fi INTERFERENCE CASES

If Bluetooth and Wi-Fi systems operate at the same time in close proximity as in the case of a mobile device then
they will interact (collide) with each other, introducing an undesired effect called interference which deteriorates the
overall performance of the wireless communication systems . The sidebands of each transmission must also be
accounted for. Interference between Bluetooth and Wi-Fi occurs when either of the following is true:
x A Wi-Fi receiver senses a Bluetooth signal at the same time when a Wi-Fi signal is being sent to it. The effect is
most pronounced when the Bluetooth signal is within the 22 MHz-wide passband of the Wi-Fi receiver.
x A Bluetooth receiver senses a Wi-Fi signal at the same time when a Bluetooth signal is being sent to it; the
effect is most pronounced when the Wi-Fi signal is within the passband of the Bluetooth receiver.

It is worthwhile to note that neither Bluetooth nor Wi-Fi was designed with specific mechanisms to combat the
interference that each creates for the other. As a fast frequency-hopping system, Bluetooth assumes that it will hop
away from occupied channels, minimizing its exposure to interference. The Wi-Fi MAC layer, which is based on the
Ethernet protocol, assumes that many stations share the same medium, and therefore, if a transmission fails, it is
because two Wi-Fi stations tried to transmit at the same time. This report examines how this assumption drives
system behavior that actually worsens the impact of Bluetooth interference.

4. ADAPTIVE BLUETOOTH SOLUTIONS

To alleviate the potential problems of interference in the ISM band, the Bluetooth SIG are discussing several
adaptive technologies [8].

4.1 Adaptive Frequency Hopping

This involves a change in Bluetooth frequency hopping sequence to allow for more flexible use in the ISM band.
Bluetooth device must hop through all 70 channels regardless of whether another technology that operates in the
ISM band is already occupying a segment of the band. It introduces some degree of intelligence into the process so
that the Bluetooth device would analyze the available spectrum and steer the transmission to those channels where
interference would not be encountered or where it would not cause interference to other devices using occupying a
portion of the band.

4.2 Transmission power control

This involves adapting transmit power of devices used in the ISM band. The reasoning behind this technique is
based on common sense. Transmitting data at a power level above the minimum needed to meet a predetermined
level of acceptable data integrity unnecessarily causes interference to other users in the band. Noting is gained by
transmitting at a higher power level above the minimum needed which would result in interference with other
devices in the area. The current Bluetooth standard calls for a poor receiver sensitivity level of -70 dBm. More
 R. Challoo et al. / Procedia Computer Science 12 (2012) 386 – 391 389

sensitive receivers would allow for a reduction in the transmit power level while maintaining an acceptable signal-
to-noise ratio which would enhance the system’s co-existence performance. Adaptive power control could also
reduce the overall power consumption by a Bluetooth device.

4.3 Adaptive Selection of Packet Type

The type of Bluetooth packets being transmitted also affects the co-existence performance. The packets can carry
various payloads depending on the number of “slots” in the packet. Packets can occupy between one to five time
slots according to Bluetooth specifications. While carrying 10 times as much data, a Bluetooth packets with five
slots will remain on a certain channel at a certain frequency five times longer than would a one-slot packet,
increasing the vulnerability of this packet to interference as well as increasing the chance with other sharing the
frequency. A transmitter is capable of dynamic packet type selection would determine would determine where and
when interference is present and adapt the Bluetooth packet type accordingly. So if a channel is acquired, more data
can be transmitted while a shorter packet type could help ameliorate the condition when interference in the
surrounding area reaches a point where packet corruption is unacceptable. Research has shown that shorter
Bluetooth packets can improve data throughput in an environment with interference.

5. 802.11 SOLUTIONS

The 802.11 solutions are also similar to the Adaptive Bluetooth techniques. As previously described in section IV,
three channels centered on channel 1, 6 and 11 are configured in figure 1 above. An 802.11 access point would
typically be assigned to a certain channel and this would not change without the intervention by the LAN
Administrator. However, dynamic channel selection would allow the access point itself to determine which channel
is best suited for communication at any time. By detecting interference on a Wi-Fi channel, a channel with high
noise content can be avoided. Multipath propagation and inter-Symbol interference are monitored and can form
basis for dynamic channel selection.

Adaptive packet fragmentation is another technique used by Wi-Fi to cope with co-existence interference. This is
because the length of 802.11b packets needed not be the maximum length for each transmission, fragmented of
shortened packets can be used to overcome the effects of coexistence interference. With shorter packet length, less
data must be retransmitted when a packet transmission fails because of interference.
Similarly, transmission power control can also minimize the interference caused to other users in the band. Here, the
optimal transmission power would be the minimum level necessary to maintain a predefined level of data integrity.

6. ZigBee SOLUTIONS

The most difficult thing when trying to remedy the interference problem between ZigBee and other devices that
share same spectrum is due to the difference in their physical layers. There are two ways for ZigBee devices to
intervene between 802.11 and Bluetooth devices that operate with it in same spectrum. The first method is
transmitting an 802.11 or a Bluetooth packet indicating that this packet would have an unusually long duration
permitting ZigBee to transmit during this period in which other 802.11 or Bluetooth device would sleep. The second
method would be the use of Request to Send (RTS) or Clear to Send (CTS) message to clear 802.11 or Bluetooth
traffic. This works on the theory that sending out a CTS message will block all 802.11 or Bluetooth devices from
transmitting for a specified period of time. Hence, the goal of the two solutions is to temporarily block out 802.11 or
Bluetooth messages for a window of time large enough that ZigBee device can successfully transmit their messages
thereby resolving interference issue [9]

7. COEXISTENCE TESTING

Since Bluetooth devices hop over 79 MHz of the ISM band, Wi-Fi devices require approximately 16 MHz of
bandwidth to operate and ZigBee has sixteen networks in the ISM Band with 2 MHz bandwidth, the possibility of
having Bluetooth, ZigBee and Wi-Fi products in same area without interference is low. Due to this interference, a
coexistence test was run with actual ZigBee, Bluetooth and Wi-Fi products to determine their level of coexistence.
390 R. Challoo et al. / Procedia Computer Science 12 (2012) 386 – 391

7.1 Testing Setup

The throughput testing was performed using a Wi-Fi access point and a station. Two laptops with Bluetooth
enabled are used as a master and slave for the Bluetooth test while two devices using ZigBee are equally used. The
test is intended to obtain the empirical data-throughput results corresponding to realistic scenarios.
A base-line test was performed to determine the maximum throughput for both the Wi-Fi and Bluetooth network.
Figure 2 shows the baseline obtained for the Wi-Fi when data was transferred from the access point to the station at
a specified distance. The result shows that the device maintains a connection speed in excess of 5.5 Mbps up to the
maximum distance of 250 feet [10].

Figure 2: Wi-Fi Baseline Throughput [10] Figure 3: Bluetooth Baseline Throughput [10] Figure 4: ZigBee Baseline Throughput [10]

The Bluetooth and ZigBee baseline throughput is shown in figure 3 and 4 respectively. Data was transferred from
the Bluetooth master to the slave with no interference. The throughput is approximately 550 Kbps at the specified
distance and 250 Kbps for ZigBee.

8. EXPERIMENTAL RESULT

The impact of Wi-Fi on an ongoing Zigbee transmission when the ZigBee network is set up on channel 11 - 14
while the Wi-Fi network is set up on channel 1 with a payload of 64-byte payload frames is very apparent. The
frame error rate as shown in figure 5 of zigbee drops to a value of 0.45 and 0.73 with channel 13 being the less
impared and 14 the most impared [11]. Hence the prensense of Wi-Fi is not totally destructive for ZigBee network.
Very similar result would be registered for other channel combinations such as Wi-Fi channel 6 and ZigBee
channels 16-19 and Wi-Fi channel 11 and ZigBee channels 21 -24. ZigBee loses a litle percentage of its frames.
Conversely, the impact of ZigBee of a Wi-Fi connection is low. Wi-Fi is practically not affected by the activity of
the ZigBee as shown in figure 6.

Figure 5: FER of ZigBee under the interference of Wi-Fi[11] Figure 6: FER of Wi-Fi under interference of ZigBee [11]

The effect of Bluetooth over the ZigBee network is not noticeable. Its Frame Error Rate is only reduled by less
than 10% [12]. Conversely, the effect of the ZigBee over the Bluetooth shows rather an unstable network, shown in
figure 7, the degradation effect is not well noticed.
 R. Challoo et al. / Procedia Computer Science 12 (2012) 386 – 391 391

Figure 7: ZigBee effect over Bluetooth [11] Figure 8: Bluetooth Effect over Wi-Fi [11]

The effect of Wi-Fi over Bluetooth shows a strong degradation of Bluetooth signal in Figure 8. Bluetooth goodput
drops from 1.12 Mbps to 0.59 Mbps for TT –RR scenario and 0.95 Mbps to 0.30 Mbps for TR –RT scenario. This
shows that the FHSS employed by the Bluetooth is less effective. As a consequence, the use of smart and adaptive
hopping patterns techniques could indeed be a major contributor towards the achievement of robust Bluetooth
connections in the vicinity of Wi-Fi devices. Wi-Fi has a great impact on both ZigBee and Bluetooth devices. The
degradation on Bluetooth occurs as soon as Wi-Fi is activated whereas it occurs only when the Wi-Fi frame rate is
increased in ZigBee.

9. CONCLUSION

To summarize, Wi-Fi devices are scarsely affected by the presence of other wireless technologies operating
concurrently such as ZigBee and Bluetooth devices Conversely, Bluetooth and ZigBee suffers conspicously from
the presence of Wi-Fi . ZigBee which was deemed to be interference free was heavily influenced by the Wi-Fi
whose spectrum covers more than the classic four ZigBee channels. Hence ZigBee cannot only rely on the
supportedly safe channels such as 15, 20, 25 and 26 [11]. The FHSS technique used by the Bluetooth did not reveal
to be very effective in contrasting Wi-Fi interference. Also, the ZigBee system is much more sensitive to the
position of the Wi-Fi transmitter than the Bluetooth. This means that while ZigBee networks can be deployed in a
shared area by having ZigBee devices placed far from Wi-Fi radios, this is not true for Bluetooth networks which
instead requires a more drastic seperation from Wi-Fi polluted areas.

REFERENCES
1. Specification of the Bluetooth System core, Version2.0 [Online]. Available: http://www.bluetooth.com.
2. I. Howitt and J. A. Guttierrez, “Low-Rate Wireless Personal Area Networks – Enabling Wireless Sensors with IEEE
802.15.4, Proc. IEEE WCNC’03, vol.3, pp. 1481-1486.
3. A. Sikora, “Coexistence of IEEE 802.15.4 with IEEE 802.11, Bluetooth and Microwave Ovens in 2.4GHz ism-Band,”
web document, http://www.ba-loerrach.de/stzedn/
4. T. Yucek and H. Arslan, “A survey of spectrum sensing algorithms for cognitive radio applications,” IEEE
Communications Surveys and Tutorials, vol. 11, no. 1, pp. 116-130, 2009.
5. A. He, K. K. Bae, T. Newman, J. Gaeddert, K. Kim, R. Menon, L. Morales-Tirrado, J. Neel, Y. Zhao, J. Reed, and W.
Tranter, “ A Survey of Artificial Intelligence for cognitive radios,” IEEE Transaction on Vehicular Technology, vol. 59,
no. 4, pp. 1578-1592, May, 2010.
6. http://en.wikipedia.org/wiki/ZigBee
7. “IEEE Standard 802.15.4-2006,”September, 2006.
8. Oren Eliezer, Matthew Shoemake, “Bluetooth and Wi-Fi co-existence schemes strive to avoid chaos” RF wireless
connectivity paper. November 2001. Pp. 66-72.
9. James H., Benjmin C., Dae-Ki C., Mario G., “ Minimizing 802.11 Interference on ZigBee Medical Sensors” Digital
Object Identifier:10.4108/ICST.BODYNETS2009.6029http://dx.doi.org/10.4108/ICST.BODYNETS2009.6029.
10. Mathew B. Shoemake “Wi-Fi (IEEE 802.11b) and Bluetooth Co-existence Issues and Solutions for the 2.4GHz ISM
Band” A white Paper of Texas Instruments. February 2001, Version 1.1
11. R.G. Garroppo, et al, “Experimental assessment of co-existence of ZigBee, Wi-Fi, and Bluetooth devices” in IEEE
2011 proceedings.
 The 5th Annual IEEE International Conference on
Cyber Technology in Automation, Control and Intelligent Systems
June 8-12, 2015, Shenyang, China

Indoor Human Tracking and State Estimation by

Fusing Environmental Sensors and Wearable
Sensors
Yanfei Li
Dept. of Systems Science
and Engineering
College of electrical engineering
Zhejiang University, Hangzhou
310027, P. R. China
Email: liyanfei@zju.edu.cn
Meiqin Liu
Dept. of Systems Science
and Engineering
College of electrical engineering
Zhejiang University, Hangzhou
310027, P. R. China
Email: liumeiqin@zju.edu.cn
Weihua Sheng
School of Electrical
and computer Engineering
Oklahoma State University, Stillwater
OK 74078 USA
Email: weihua.sheng@okstate.edu

Abstract—In this paper, we aim to track a human’s indoor
location while estimating his/her behavioral state. We use en-
vironmental PIR sensors to localize the individual and assume
that the position of the individual obeys the bivariate Gaussian
distribution, which can be used to improve the tracking accuracy.
At the same time we introduce a wearable acceleration sensor
to estimate the human’s state. The hardware setup consists of
two types of sensor: passive infrared sensors and a three-axis
acceleration sensor with built-in bluetooth communication. By
comparing the tracking performance between the PIR sensor-
based algorithm and the fusion-based algorithm, we find that
the latter has less errors. This indoor localization system can be
used in future smart homes.
I. INTRODUCTION
In recent years, youngsters become increasingly busy and
have no time to take care of their elder parents. On the other
hand, the elderly population is growing rapidly. Some studies
indicate that the elderly population had increased to almost
810 million in 2012, and predict that in 2050, the number of
aged people (60 and above) is about to reach a staggering 2
billion [1][2]. So elderly care has become a burning problem.
Rather than moving to expensive adult care or health-care
facility, elders would prefer to stay at their own comfortable
home [3]. In order to improve the quality of elders’ life
while young people having hectic work at daytime, under
the assistance of some smart home intelligent devices, elders
would be encouraged to maintain independent in their life. In
order to accomplish this we should firstly know the location
of the elders [4], tracking their moving path and estimate
whether they walk out. GPS (Global Positioning System) plays
an important role in location tracking outdoors. But since the
GPS signal gets weak inside the building, it cannot be used
in indoor environments [5]. Indoor human tracking methods
can be divided into terminal based and non-terminal based.
Terminal based approaches use RFID tags and other mobile
devices carried along with the individual to be tracked. Non-
terminal based approaches use ultra sound, passive infrared
sensors, smart floors and cameras for location tracking. Cam-
eras have high accuracy but they raise privacy concerns and
are not acceptable to many in their daily life. Carrying too
many terminal devices would be very uncomfortable, so we
should use the least terminal devices while at the same time
reach the maximum improve accuracy goal. Here we introduce
a single terminal device, a wearable acceleration sensor, to
obtain the human motion information. In this paper we adopt
a method that fuses data from PIR sensors with data from
the wearable acceleration sensor. The PIR sensors used in
this thesis are “ON/OFF” sensor. Hence it is suitable for the
application for indoor localization. We distribute multiple PIR
sensors throughout the home in a certain pattern. These sensors
along with the acceleration sensor are used to estimate whether
he/she stands still in the test area or walks out of the test area
in addition to tracking the individual’s indoor location [6].
This paper is organized as follows. Section II presents
the architecture of the indoor localization system. Section III
describes the methodology of human tracking through the
fusion of PIR sensors and a wearable acceleration sensor.
Section IV gives the result of the experiments. Finally, Section
V presents the conclusion and the future work.
II. H ARDWARE P LATFORM
A. Overall system
The hardware platform used in this paper consists of multi-
ple PIR sensor nodes and a three-axis acceleration sensor with
bluetooth low energy (BLE) communication. The PIR sensor
node is shown in Fig.1, and it includes the following hardware
parts.
• Arduino Mega 2560
• Xbee shield
• Xbee module
• Panasonic PIR sensor
The Arduino Mega 2560 is a microcontroller board based
on the ATmega2560 microcontroller, and it is the brain of
the sensor node. It has 54 digital input/output pins, 16 analog
inputs, 4 UARTs (hardware serial ports), a 16 MHz crystal

978-1-4799-8730-6/15/$31.00 © 2015 IEEE 1468


Fig. 1. PIR sensor node
Fig. 2. The three-axis acceleration sensor.
oscillator, a USB connection, a power jack, an ICSP header,
and a reset button. An XBee shield along with an Xbee module
is placed on the Arduino board. The Xbee module is used to
transmit the sensor data using Zigbee protocol to the gateway
node.
The acceleration sensor node is shown in Fig.2, and it
includes:
• CC2540 Keyfob board
• CC2540 USB dongle
The CC2540 Keyfob is designed to act as a Peripheral
Device(BLE Slave). The keyfob operates on a single CR2032
coin cell battery and includes a two-colored LED, a buzzer,
an accelerometer and two buttons. The keyfob uses SPI to
interface with the three-axis accelerometer, CMA3000d. The
CC2540 USB Dongle can be used to emulate any Bluetooth
low energy behavior but usually acts as a Central Device(BLE
Master). It connects to Windows PC’s USB port, and is pre-
loaded with necessary software to run the PC application
BTool.
The PIR sensor network is used to detect the location and
track the elders. The network consists of eight PIR sensor
nodes. PIR sensors can tell us the approximate location of
a human living in the environment, but we could not know
whether he walk out of the test area. In order to solve this
problem, we introduce the three-axis acceleration sensor, based
on which we can detect the steps when the individual is
1469
Fig. 3. Circuit diagram of PIR sensor board
Fig. 4. The computation of the height of the paper cylinder
walking and fuse the two types of sensor to improve the
tracking accuracy and estimate whether the individual is in
or out.
B. The setup of the PIR sensor
Fig.3 is the circuit diagram of the PIR sensor board. The
motion detection LED is turned on whenever the sensor detects
motion in its field of view. As the sensor has a very large field
of view, we attach a paper cylinder to it to adjust the field
of view. By doing this we could change the sensor area by
increasing or decreasing the height of the paper cylinder. In
this paper, we set the radius equal to 2.2 centimeter. The radius
of the detection range equal to 1.5 meter, and the distance from
the sensor to the ground equal to 2.7 meter, so we can calculate
the height of the paper cylinder is equal to 4.0 centimeter.
The computation method of the height of the paper cylinder
is demonstrated in Fig.4.
The PIR circuit board is fixed to the Arduino board. The
Xbee shield along with the Xbee module is fitted to the
Arduino board. One LED is used to indicate the motion
detection. The Arduino is programmed using Arduino IDE.
The firmware on the Arduino reads data from the sensor and
transmits it to the gateway node through the Xbee module,
that is to say, if the PIR sensor detects the passive infrared
signal, the Xbee module transmits “ON”, otherwise it transmits
“OFF” [7][8]. When motion is detected, the sensor outputs a
digital high and an encoded data is transmitted through the
Xbee module, and the LED is turned on. The encoded data
contains the sensor ID and the status of the motion. When the
motion stops, the sensor outputs a low voltage and another
encoded data is transmitted to the gateway node, and the LED
 Fig. 5. The integrated framework of the PIR network
is turned off. The individual is considered to be in the sensor’s
field of view when the LED is turned on and when the LED
is off it indicates that the individual stops motion or walks out
of the sensor’s field of view. It is worth noting that there is
a sensor output delay when the individual enters or exits the
sensor’s field of view [5]. The whole PIR sensor network is
mounted on the ceiling of our lab using a rectangular frame
made of PVC tubes. The mounting of the PIR sensor network
is shown in Fig.5, from which we can see eight PIR sensor
nodes. We place three PIR sensor nodes on each of the long
sides of the rectangle, and the spacing between two adjacent
sensor is 1.5 meter. One sensor node is placed in the middle of
the each short side of the rectangle. The power to the sensors
is controlled by a switch on the celling in the center of the
rectangle. The power is provided to every PIR sensor node
through a USB hub.
C. The acceleration sensor
The acceleration sensor is based on the CC2540 mini kit,
which is a ultra-low power consumption system-on-chip, and
integrates the microcontroller, the host and the application on
one device. The CC2540 mini kit combines an excellent RF
transceiver with an industry-standard enhanced 8051 micro-
controller, which includes an analog and digital peripherals
sensor, built-in programmable flash memory, precise radio
frequency signal strength indicator, full-speed USB 2.0 in-
terface, built-in AES-128 encryption engine. After setting the
parameters using BTool, we enable accelerometer notification.
The first characteristic of the accelerometer service has both
read and write permissions and the accelerometer is enabled
by writing a characteristic value handle “0x0034” and Hex
value “01:00”. Once the accelerometer is enabled, each axis
can be configured to send notifications by writing Hex value
“01:00” to the characteristic configuration for x-axis,y-axis and
z-axis with its characteristic value handle “0x003B”,“0x003F”
and “0x0043”, respectively. Then CC2540 USB Dongle can
receive three-axis signal from CC2540 Key-fob [9]. It is worth
noting that every time the USB Dongle is unplugged, the
connection between the Dongle and the Key-fob will be broken
and we must connect them and enable the acceleration again.
1470
Fig. 6. Flowchart of PIR sensor and acceleration sensor
III. T HE F USION A LGORITHM
A. Overall approach
In this section, we use the sensor fusion algorithm to
improve tracking accuracy and estimate the individual’s state
[10]. Our algorithm has two parts. Firstly, we use the PIR sen-
sor network to estimate the general area where the individual
may be in, and at the same time we use the distribution of the
furniture to improve the accuracy of the estimation. Secondly,
we use the acceleration sensor placed on the foot to determine
whether the individual stands still or walks out. The flowchart
of the sensor fusion algorithm is shown in Fig.6 [5].
B. Tracking and state estimation method
In practical experiment, we have a rectangle test area with
a size of 4.5m by 4.6m on floor, and there are 8 PIR sensor
nodes, of which the detection area distribution is presented
in Fig.7. The PIR sensors divided the experiment area into
31 areas [11][12], marked with a-s,A,a1-a4,B,b1-b4, where A
and B stand for two areas, because they are too small to be
missed easily. The individual moves in different sensors’ field
of view and these sensors will detect the infrared signal, and
transmit the data through Xbee module in the end device to
the coordinate Xbee module. At the same time the LED is
turned on, then the program reads the USB serial to obtain
positional information of the individual [13][14]. In the test
area, there are a bookcase, a desk and a chair where one can
sit down.
With this method we can obtain a rough track [11]. In every
area, assuming that the probability of human location is subject
to bivariate Gaussian distribution [15],i.e.,
1 1 T
Φ=  exp (− (x − μix )Vi−1 (y − μiy ) ) (1)
2 2
(det Vi )(2π)
 4.5 4.5
s r q p o
4 4

3.5 3.5
j l m
k Book−
n 3

y−axis coordinate/m
3
case
2.5
a1 a3 b1 b3 2.5
i t B
A
2 a2 a4 b2 b4 2

1.5 g e
Desk 1.5
h f
1
1

0.5
0.5 b
a c d Chair
0
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0 1 2 3 4
x−axis coordinate/m

(a) Tracking based on PIR sensor

Fig. 7. The detection area divided by PIR sensor
4.5

4
Where, (μix , μiy ) is the mean of the distribution function. For
our detection, it is the coordinate of the center point of some D
3.5

PIR sensors’ overlapping field of view area, and Vi is the 3

y−axis coordinate/m
stop2
covariance matrix, as the coordinate of X-axis and Y-axis is 2.5
uncorrelated, so
  2
2
σxi 0 stop1
Vi = 2 (2) 1.5
0 σyi C
1

In our experiment, the coordinate of X-axis and Y-axis is 0.5

equal to zero when the serial could not receive data transmit
from the PIR sensors within α seconds, that is to say, there is 0
0 1 2 3 4

no PIR sensor being activated during α seconds, where α can

x−axis coordinate/m

be adjusted. If the time between two adjacent data is less than (b) Tracking based on PIR sensor and acceleration sensor
α, we store the data from the PIR sensor in an array, otherwise
Fig. 8. The first tracking trajectory when α = 0.3second(the blue trajectory
we process the data in the array. Generally, if α is too large, indicates the real rough path, and the black trajectory indicates pursuit path)
realtimeness of the tracking will deteriorate and we may miss
some smaller areas which the individual may pass through. If
α is too small, the information of some PIR sensor nodes will second path is following the area marked with the alphabet
be lost. Consequently it is reasonable to set α between 0.3 “a,h,i,A(a1,a2,a3,a4),t,b1,m,o-s” as shown in Fig.7.
second and 0.5 second. When the coordinate of X-axis and Y-
axis is equal to zero it indicates that he/she stands still or out IV. E XPERIMENTAL R ESULT
of the test area. In order to classify the state of the individual, A. Sensor reading
the acceleration sensor attached to the right foot will be used. In our experiment, there are two kinds of sensor, so the
If the PIR sensors do not detect anything, or the coordinate of two serial ports (one is connected to the coordinate XBee
X-axis and Y-axis is equal to zero, the acceleration sensor will node, another is connected to CC2540 USB Dongle) must be
be read and we can use the acceleration data to predict the state read simultaneously. If we read the two serials in sequence, it
of the individual. When the acceleration data equal to zero it will lead to some data loss and affect the real-time tracking
indicates that the individual stands still somewhere in the test performance. In order to solve this problem, after many trials,
area, otherwise it indicates that the individual walks out of the an inquiry mode is found. In this mode, we read the CC2540
test area. At the same time we can use the location information USB Dongle when the PIR sensors have no data to transmit.
of the previous time to predict whether the individual walks Processing the acceleration sensor data as fast as possible is
out of the test area. For example, if the individual is in the area helpful to avoid data loss and improving the real-time tracking
marked with “a1” in Fig.7 now, it is impossible that he/she performance.
could walk out of the test area while not walking through other
edge areas in the next moment, which drastically reduces the B. Tracking walking path
chance of false positives. In this paper, we introduce two kinds The experiment includes two kinds of walking path. The
of walking path. The first path is from the area marked with first path is shown in Fig.8, where the blue trajectory indicates
“a” to the area marked with “s” in a sequence of alphabet. The the real path and the black trajectory indicates the estimated

1471
 4.5 4.5

4 4

3.5 3.5

3 3
y−axis coordinate/m

y−axis coordinate/m
2.5 2.5

2 2

1.5 1.5

1 1

0.5 0.5

0 0
0 1 2 3 4 0 1 2 3 4
x−axis coordinate/m x−axis coordinate/m

(a) Tracking based on PIR sensor (a) Tracking based on PIR sensor

C
4.5 4.5

4 4

3.5 3.5

3 3

y−axis coordinate/m
y−axis coordinate/m

2.5 2.5
C
2 2

1.5 1.5
stop

1 1

0.5 0.5

0 0
0 1 2 3 4 0 1 2 3 4
x−axis coordinate/m x−axis coordinate/m

(b) Tracking based on PIR sensor and acceleration sensor (b) Tracking based on PIR sensor and acceleration sensor

Fig. 9. The second tracking trajectory when α = 0.1second(the blue trajectory Fig. 10. The second tracking trajectory when α = 0.5second (the blue
indicates the real rough path, and black trajectory indicates pursuit path) trajectory indicates the real rough path, and the black trajectory indicates
pursuit path)

path. The second path is shown in Fig.9 and Fig.10, the

blue trajectory indicates the real path and the black trajectory
indicates the estimated path. The difference between Fig.9 and
Fig.10 is that they have different α, and obviously, different
α leads to different tracking accuracies.
C. Comparison
Fig.8(a) presents a tracking path only based on the PIR sen-
sor. We assume that if PIR sensor nodes detect the individual
we set the position coordinate of the individual equal to the
coordinate of the center point of the overlapping field of view
area. Fig.8(b) depicts the tracking path based on the fusion
of PIR sensor nodes and acceleration sensor. In this method,
we assume that the position coordinate of the individual in
the overlapping field of view area is subject to the bivariate
Gaussian distribution and the mean of Gaussian distribution
is the coordinate of the center point of the overlapping field
of view area, the variance changes among 0.1, 0.2 and 0.25
according to the size of different detection overlapping area.
Fig.8 clearly demonstrates that the accuracy of tracking after
fusion significantly improves and in Fig.8(b), there are two
stop points marked with ”stop1” and ”stop2” and two points
where the individual exits and enters the test area marked with
“C” and “D”.
Fig.9 and Fig.10 present the improvement of tracking ac-
curacy. They also demonstrate the same path with different
α. α in Fig.9 is equal to 0.1 and in Fig.10 is equal to 0.5.
Comparing the two figures, we can find that there are some
PIR sensor data loss if α is too small, and this leads to
a decrease in the accuracy of tracking, as demonstrated in
Fig.9. The accuracy of tracking is improved greatly in Fig.10
as we set the parameter α much greater, and overcome the
poor performance of losing some PIR sensor node data. It
is worth noting that we should not set the parameter α too
large, because if α is too large we will miss some smaller
PIR sensor nodes’ overlapping field of view area which the
individual may pass through.
Fig.9(b) demonstrates that in the small oval area marked
with “C” the individual exits the test area at the right point
and enters the test area at the left point, and there is no point
where the individual stops. While Fig.10(b) demonstrates that

1472
the individual exits and enters the test area at the point “C”.
At the same, it also demonstrates that the individual stops at
the point marked with “stop”.
V. C ONCLUSION AND FUTURE WORK
This paper introduced the indoor localization setup which
includes the PIR sensor nodes and the acceleration sensor,
and the fusion method of PIR sensor and acceleration sensor.
In our experiment, we have only eight PIR sensor nodes and
we can only detect the rough area where the individual walks
through instead of an accurate walking path. Some measures
have been taken to improve the tracking accuracy, such as
using the furniture distribution to limit the walking path and
using the coordinate obtained from the previous time and the
bivariate Gaussian distribution to decrease the tracking error.
Consequently, comparing the performance of the tracking only
based on PIR sensor with tracking based on the fusion of
the PIR sensor and the acceleration sensor, the latter had
better tracking accuracy. Besides, by combining data read from
the acceleration sensor with information from the PIR sensor
nodes, we can know whether the individual walks out of the
whole test area or not.
More research will be conducted to extend the current
proposed method. Smart home robot will be introduced in our
research to assist elders when they have abnormal behaviors.
In the future, the robot will be used to built the map of
complicated and dynamic environments. Combining the map
with the individual’s location which will be used as an initial
sensing modality for the robot [16], the robot can find and
assist elders in the shortest time.
[7] N. Shrivastava, R. Mudumbai, U. Madhow, and S. Suri, “Target tracking
with binary proximity sensors,” ACM Trans.Sensor Netw., vol. 5, no. 4,
pp. 1–33, 2009.
[8] P. Bahl and V. Padmanabhan, “Radar:an in-building rf-based user
location and tracking system,” in Proc.IEEE INFOCOM, pp. 775–784,
2000.
[9] “Bluetooth low energy cc2540/41 mini development kit users guide.”
[10] R. C. Luo,and O. Chen, “Wireless and pyroelectric sensory fusion sys-
tem for indoor huaman/robot localization and monitoring.” IEEE/ASME
Transactions on mechatronics, vol. 18, no. 3, pp. 845–853, June. 2013.
[11] K. N. Ha, K. C. Lee, and S. Lee, “Development of pir sensor based
indoor location detection system for smart home,” in SICE-ICASE
International Jiont Conference., Bexco,Busan,Korea, Oct 18-21, 2006,
pp. 2162–2167.
[12] S. Lee, K. N. Ha, and K. C. Lee, “A pyroelectric infrared sensor-based
indoor location-aware system for the smart home.” IEEE Transactions
on Consumer Electronics., vol. 52, no. 4, pp. 1311–1317, 2006.
[13] N. Patwari, A. O. Hero, M. Perkin, N. Correal, and R. J. O’Dea,
“Relative location estimation in wireless sensor networks,” IEEE Tran-
s.Sig.Proc., vol. 51, no. 8, pp. 2137–2148, Aug. 2003.
[14] N. Patwari, R. J. O’Dea, and Y. Wang, “Relative location in wireless
networks,” in Proc.IEEE VTC Conf., vol. 2, pp. 1149–1153, May. 2001.
[15] H. M. La and W. H. Sheng, “Distributed sensor fusion for scalar
field mapping using mobile sensor networks,” IEEE Transactions on
cybernetics, vol. 43, no. 2, April. 2013.
[16] K. C. Lee, and H. H. Lee, “Network-based fire-detection system via
controller area network for smart home automation,” IEEE Transactions
on Consumer Electronics, vol. 50, no. 4, pp. 1093–1100, Nov. 2004.

ACKNOWLEDGMENT
This work was supported in part by the Nation-
al Natural Science Foundation of China under Grants
61328302,61222310,and 61174142, the Zhejiang Provin-
cial Natural Science Foundation of China under Grant
LZ14F030002,the Specialized Research Fund for the Doc-
toral Program of Higher Education of China (SRFDP) un-
der Grants 20120101110115 and 20130101110109, the Fun-
damental Research Funds for the Central Universities un-
der Grant 2014XZZX003-12, and the ASFC under Grant
20132076002.

R EFERENCES
[1] “10 facts on ageing and the life course,” World Health Organization,
May 26 2014.
[2] “ageing in the twenty- first century: A celebration and a challenge,”
UNFPA (United Nations Population Fund) and HelpAge International,
2012.
[3] “elderly care,” Wikipedia, May 19 2014.
[4] H. Koyuncu and S. H. Yan, “A survey indoor positioning and object
locating systems,” IJCSNS Int. J.Comput.Sci. Netw.Security, vol. 10,
no. 5, pp. 121–128, May. 2010.
[5] D. C. Kallur, “Human localization and activity recognition using dis-
tribute motion sensors,” Master’s thesis, Oklahoma State University, July
2014.
[6] Y. Gu, A. Lo and I. Niemegeers, “A survey indoor positioning system for
wireless personal network,” IEEE Commun.Surveys Tutorials, vol. 11,
no. 1, pp. 13–32, First Quarter. 2009.

1473
K. C. Lee and H.-H. Lee: Network-based Fire-Detection System via Controller Area Network for Smart Home Automation
Network-based Fire-Detection System via Controller Area
Network for Smart Home Automation
Kyung Chang Lee, Associate Member, IEEE, Hong-Hee Lee, Member, IEEE
Abstract — This paper presents a network-based fire-
detection system via the controller area network (CAN) to
evaluate the feasibility of using such a home automation
protocol in a smart home. In general, a conventional fire-
detection system has several shortcomings, such as weakness
to noise, because it uses an analog transmission with 4 – 20
mA current lines. Hence, as an alternative to a conventional
system, this paper describes the structure of a CAN-based fire-
detection system and the design method of a CAN
communication network. The performance of the proposed
system is evaluated through experimental tests. The CAN has
several advantages, such as low cost and ease of
implementation, as compared to other low layers of the
BACNet, such as Ethernet or ARCNET. Therefore, if the CAN
is selected for the low layer of the BACNet, a home
automation system can be implemented more effectively1.
Index Terms — network-based fire-detection system,
controller area network (CAN), home automation system,
home network system, intelligent building.
I. INTORDUCTION
Recently constructed buildings tend to be “intelligent” to
enhance the convenience and safety of occupants [1][2]. Thus,
the requirements for home networks and automation systems
have increased with the demand for smart homes [3]. To
satisfy the needs of occupants, home appliances such as
refrigerators or microwave ovens, multimedia devices such as
televisions or audio systems, and Internet devices such as PCs
have been included in intelligent buildings, as shown in Fig. 1.
In smart homes, appliances are connected to the home network
so that it becomes possible to control and monitor the
appliances from inside the house or remotely using a cellular
telephone or PDA. To realize a home network system, several
standards, such as Echonet, Konnex, LnCP, and LonWorks,
are being developed by standards organizations and
corporations [4].
In addtition, home automation systems, such as HVAC
(heating, ventilation, and cooling), lighting, anti-crime, and
fire-detection systems are being used to improve convenience
1
This work was supported (in part) by the Korea Science and Engineering
Foundation (KOSEF) through the Network-based Automation Research
Center (NARC) at University of Ulsan.
H.H. Lee is with the School of Electric-Electronic Information System
Engineering, University of Ulsan, Ulsan 680-749, Korea (corresponding
author to provide phone: 82-; fax: 82-; e-mail: hhlee@main.ulsan.ac.kr).
K.C. Lee is with the Network-based Automation Research Center,
University of Ulsan, Ulsan 680-749, Korea (e-mail: gclee@pnu.edu).
Contributed Paper
Manuscript received July 29, 2004 0098 3063/04/$20.00 © 2004 IEEE
1093
and safety for occupants. Generally, in a conventional home
automation system, home equipments, such as switches,
valves, or fire detectors, are directly connected to controllers
for the HAVC or fire-detection systems. However, because of
the analog transmission method, which uses a 4 – 20 mA
current, conventional fire-detection systems detect a fire as
occurring when the current received from a fire detector
exceeds a predefined threshold. Hence, the system has
disadvantages such as a weakness to noise of various forms,
including impulses or short-circuits, and a lack of awareness of
the actual location of a fire. Recently, in order to solve these
problems, research on fire-detection systems that use digital or
wireless transmission instead of analog transmission has been
conducted [5][6]. In parallel with this, research is also being
performed on communication protocols, such as the BACNet
or LonWorks [7].
Recently, researchers have proposed applying the controller
area network (CAN) [8] to the home network protocol, and
attention has increasingly been centered on the CAN in
automation (CiA) [9]-[11]. Especially, the price of a CAN
microcontroller, which can be integrated into one-chip by
semiconductor manufacturing companies, is low and many
CAN development tools are coming onto the market. Also,
many application programs had been developed in various
fields such as automotive, robotic, and industrial automation
systems. Therefore, a home automation system using the CAN
be more easily implemented, because realization at the sensor
or actuator level is relatively straightforward as compared to a
system using an Ethernet that is the lower layer of the
BACNet.
To evaluate the feasibility of the CAN for home automation
protocols, this paper introduces a network-based fire-detection
system via the CAN. The paper is organized into five sections,
including this introduction. Section II gives the architecture of
the network-based fire-detection system. Section III presents
the design method of the CAN-based fire-detection system,
and verifies that a CAN communication network satisfies the
real-time requirements of a home automation system. Section
IV describes an experimental testbed that was implemented to
verify the feasibility of the proposed system. Finally, a
summary and the conclusions are presented in Section V.
II. ARCHITECTURE OF NETWORK-BASED FIRE-DETECTION
SYSTEM
A. Conventional fire-detection system
Fig. 2(a) shows the architecture of a conventional fire-
detection system. As shown in the figure, fire detectors (e.g.,
1094
mobile computer
Internet
home gateway
IT network
cellular phone, PDA, etc.
Fig. 1. Schematic diagram of a home network system.
smoke detector, heat detector, gas detector, etc.) and actuators
(e.g., guide light, fire wall, sprinkler, smoke ventilator, etc.)
are connected to a receiver via a dedicated analog signal line
with a current of 4 – 20 mA. In particular, several fire
detectors that guard the same area are connected to a single
signal line. Fig. 2(b) shows the connection method used
between the receiver and the fire detector in a conventional
fire-detection system. As shown in the figure, the receiver
perceives a fire based on the fact that the current increases in
the corresponding analog connection line when a fire occurs.
That is, the fire detector that perceives the fire causes the
analog line to short-circuit so that the current increases. At this
time, when the receiver detects that the current in a
corresponding line has increased, it relays this information
using a bell or a signal lamp to indicate that a fire has occurred
in the corresponding area.
Generally, the current in an analog line will be increased if
any one of the fire detectors connected to that line perceives a
fire in the conventional fire-detection system. Hence, the
receiver cannot distinguish which detector connected in a
bell
B
D D ... D in the transmission medium. In addition, the number of false
sign ramp receiver fire detector
...
D D ... D PC based receiver
(a) Basic structure of a traditional fire-detection system
4-20mA current
receiver terminator
fire detector
(b) Connection method of a traditional fire-detection system
Fig. 2. Basic structure of a traditional fire-detection system.
IEEE Transactions on Consumer Electronics, Vol. 50, No. 4, NOVEMBER 2004
home network system home automation system
home appliance
HVAC control
control network
backbone network
AV network
H
IT audio & video lighting control fire control
single analog line has sensed a fire, so the location where a fire
has occurred can not be indentified. Furthermore, because the
receiver recognizes that a fire has occurred only by the
increase in the current above a specific threshold, a
conventional fire-detection system cannot process data such as
the quantity of smoke or the degree of heat detected by the fire
detectors. If there is a breakdown in the analog line owing to
deterioration or corrosion before a fire occurs, the receiver will
neither recognize the problem nor detect the fire. Therefore,
there is a very high probability that a conventional fire-
detection system may misfunction.
B. Network-based fire-detection system
Fig. 3 shows the architecture of a network-based fire-
detection system that is able to overcome the shortcomings of
a conventional fire-detection system. As shown in the figure, a
fire detector, an actuator, a bell, and a display device are
connected by a shared transmission medium, and information
is exchanged using digital communication. Using this
connection method, the receiver can identify which fire
detector senses a fire because each fire detector has its own
terminator address. Also, because the receiver periodically examines the
state of the fire detectors, it can recognize a breakdown in the
system, such as the failure of a fire detector or an open circuit
alarms is less than with conventional systems because analog
data such as the quantity of smoke and the amount of heat
network
B D D A …
sign ramp bell fire detector actuator
Fig. 3. Structure of a network-based fire-detection system.
K. C. Lee and H.-H. Lee: Network-based Fire-Detection System via Controller Area Network for Smart Home Automation
measured by each fire detector can be sent to the receiver. It is
straightforward to apply such a system to an intelligent fire
detection system and to apply a reasoning algorithm, because
the receiver can make use of the digitized detection values of
the smoke and heat from multiple fire detectors installed in the
same guard area. As changes in the detection signals can be
calculated and compared, assessment of risk can be improved
[12].
Besides, a PC-based receiver can be used in a network-
based fire-detection system for greater convenience. Generally,
it is easier to install and maintain a PC-based receiver than a
conventional dedicated receiver. Since man-machine interface
(MMI) technology can be applied to a PC-based receiver, it is
straightforward to create a user interface and integrate the fire-
detection system into a home network system. Recently,
several protocols, such as BACNet, LonWorks, and Bluetooth,
have been developed to implement network-based fire-
detection systems in intelligent buildings [5][6].
III. DESIGN OF A CAN-BASED FIRE-DETECTION SYSTEM
A. Overview of the CAN protocol
CAN 2.0B is a network protocol that was specifically
developed for connecting the sensors, actuators, and ECU’s of
a vehicle. CAN 2.0B supports data rates from 5 kbps to 1
Mbps, which allows the CAN network to be used to share
status information and for real-time control. The network
topology can be either a linear bus or a star.
CAN 2.0B has the following properties:
z distributed medium access control, meaning that each
device has the same privilege to use the shared medium.
z contention-based and nondestructive bus access,
indicating the access to the bus has to be obtained
through some type of competition but with no loss of
network capacity due to the contention.
no
Stop
the transmission
no
Arbitration yes
won?
yes
the transmission
Fig. 4. Flowchart of the CAN protocol.
1095
content-based addressing, implying that each message
z
packet has a unique identifier according to its content.
z cyclic redundancy check for error detection, and
z error confinement to block any adverse effects of a
network component failure.
A device on the network can transmit its message whenever
the network is idle. When the network is busy, the packet
transmission has to be delayed until the on-going transmission
has finished. Because electronic signals on the network have a
finite speed of propagation, there is always the possibility that
multiple devices will start their transmissions within a short
interval. This situation, called message collision, is resolved by
comparing the identifiers of the messages involved; that is, the
message with the lowest identifier value wins the right to use
the network, while the other devices must stop their
transmissions immediately. As the identifier is located at the
beginning of the packets and the electronic signal for zero is
designed to overwrite the signal for one, the message with the
lowest identifier value finishes the contention with no damage
to the packet. The other devices will try to resend their packets
when the first transmission has finished. This arbitration
procedure is shown in Fig. 4.
B. Design of a CAN-based fire-detection system
This section presents the design methodology of a CAN for
a fire-detection system. Fig. 5(a) shows the frame exchange
between the receiver and the fire detector during the normal
state. In the figure, the receiver inspects the state of the fire
detectors by sending a poll frame to each fire detector and
actuator in the poll list once every TP (polling period). When it
receives the poll frame, each fire detector and actuator
transmits a status frame that includes its own fire detection
value back to the receiver. Thus, the receiver can periodically
perceive the state of a fire-detection system such as a fire
detector failure or a breakdown such as an open circuit.
Power on
yes no
Is the bus free?
Is a message Check
transmitted? the identifier
yes
Transmit
the message no
Receive a
transmitted
message?
Collision
occurred? yes
no
Continue Receive
the message
1096
receiver detector/actuator
poll frame
TP status frame
(a) Exchange of the status frame
receiver detector/actuator
sign of fire
TA
fire
alarm frame
TF
extinction frame extinction
(b) Exchange of the fire alarm and extinction frame
Fig. 5. Frame definition in a CAN-based fire-detection system.
Fig. 5(b) shows the frame exchange between the receiver
and a fire detector when a fire occurs. In the figure, the
detector measures values such as the amount of heat, quantity
of smoke, or quantity of gas, and thus detects the signs of a
fire. It then transmits an alarm frame that includes the detected
values to the receiver once every TA (alarm period). When the
receiver receives an alarm frame, it analyzes the real-time fire
information from multiple fire detectors located in the same
area to judge whether a fire has occurred. If the fire detection
values exceed a predefined threshold, the receiver considers a
fire to have occurred and transmits an extinction frame to the
corresponding actuators to put out the fire. The fire detectors
stop transmitting the alarm frame when they receive the
extinction frame. Here, the response time, TF, is the elapsed
time between the transmission of the alarm frame by the fire
detector and the reception of the extinction frame by the
actuator.
Using these methods, a network-based fire-detection system
will be superior to the conventional fire detectors, because the
receiver can synthetically judge whether a fire has occurred
using the directly received fire detection data from multiple
fire detectors. However, to satisfy the real-time requirements
28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10
frame fire level block
ID ID ID
Fig. 6. ID allocation method in a CAN-based fire-detection system.
IEEE Transactions on Consumer Electronics, Vol. 50, No. 4, NOVEMBER 2004
of a CAN-based fire-detection system, the alarm frame and
extinction frame must be transmitted earlier than other frames.
In order to guarantee that a network-based fire-detection
system meets the above requirement, we introduced an ID
allocation method, shown in Fig. 6. As shown in the figure, a
29-bit ID field for the CAN is organized into five ID sub-fields
to classify the four types of frame. The first portion of the
frame ID defines the type of frame. Here, to ensure that the
extinction frame is transmitted before any other frames, the
value of the extinction, alarm, poll, and status frames are set to
00, 01, 10, and 11, respectively. Fig. 7(a) shows an example of
the transmission order when several frames with different
frame IDs are generated simultaneously. In the figure, the
extinction frame of the receiver (frame ID of 00) is transmitted
before any other frames, as determined by a feature of the
CSMA/NBA algorithm used in the CAN. The alarm frame B
of a fire detector B (frame ID of 01) is then transmitted before
the status frame of a fire detector A (frame ID of 11).
The second portion of the frame ID defines the fire level,
which is used to ensure that older frames are transmitted
before newer frames. The initial fire level ID value is set to
1111. Fig. 7(b) shows an example of how the fire level ID
changes. Suppose that two frames with the same frame ID of
00 are simultaneously generated in fire detectors B and C, and
that the unique ID of the fire detector B is higher than that of a
fire detector C. At this point, the other IDs (block, sensor type,
and unique IDs) of the two frames are compared, and the
frame of the fire detector B, which has a higher priority, is
transmitted first. While, the frame of the fire detector C fails in
transmission due to a loss of competition, and then its fire
level ID value is decreased from 1111 to 1110 so that its
priority increases. Now suppose that another frame with the
same frame ID of 00 is generated in the fire detector A, before
the fire detector B starts its transmission procedure. The frame
value of the fire detector A is 001111, and that of the fire
detector C is 001110. Hence, the frame of the fire detector C is
transmitted before that of the fire detector A. Finally, the frame
of the fire detector A is transmitted. Using this logic, when
several fire detectors recognize signs of fire and transmit alarm
frames simultaneously, each frame will be transmitted to the
receiver in the correct order with a minimum amount of delay,
and the receiver can correctly identify the location of the fire.
The third portion of the frame ID is the block, which is used
to divide the fire sensors and actuators into groups. For
example, if a fire occurs in a certain area, this field is used to
transmit an extinction frame to all of the actuators that are
located in that area to put out the fire. The fourth portion of the
frame ID, the sensor type, is used to identify the types of fire
9 8 7 6 5 4 3 2 1 0
sensor type unique
ID ID
K. C. Lee and H.-H. Lee: Network-based Fire-Detection System via Controller Area Network for Smart Home Automation 1097

success
00 --
receiver

detector A
11 -- failure 11 -- failure

detector B 01 -- failure 01 -- success

(a) Example of contention owing to the frame ID

detector A 01 1111 -- failure

detector B
01 1111 -- success

detector C 01 1111 -- failure 01 1110 -- success

(b) Example of contention owing to the frame and the fire level IDs
Fig. 7. Example of contention owing to ID.

detector and actuators, such as smoke sensors, heat sensors,

gas sensors, sprinklers, fire walls, etc. Finally, the last portion
of the frame ID is a unique value that is assigned to each fire
detector and actuator.
C. Maximum transmission delay of a CAN-based fire-
detection system
Since a frame with a higher priority ID should be
transmitted before the other frames in the CAN, the lowest
priority ID may be transmitted after the maximum allowable
delay, which is defined by the network designer. Therefore, to
guarantee the performance of a CAN-based fire-detection
system, a polling cycle, TP, and an alarm cycle, TA, must be
selected appropriately so that the delay of any frame is less
than the maximum allowable delay.
In order to select TP and TA, the worst-case transmission
time (Cm) of the CAN with a 29-bit ID field can be calculated
as follows [13]-[15]:
  54 + 8s m   (1)
C m =    + 67 + 8s m τ bit
 4   which is defined as the interval between when sensors detect a
where τbit, which is the bit time, is the inverse of the
transmission speed [16]. For example, if the transmission
speed of the CAN is 500 kbps, the bit time is 2 µs. Also, sm is
the byte length of the CAN message. The first part of sm is the
size of the worst case bit value, the second part of sm is the
overhead length of the CAN message, and the last part of sm is
the bit length of the CAN message.
In the worst case that alarm frames are received from all fire
detectors and extinction frames are transmitted to all actuators,
all frames must be transmitted within the maximum allowable
delay to guarantee the performance of the CAN-based fire-
detection system. Hence, TP and TA should be selected as
follows:
Nd Na Nd + Na
(TP , TA ) > ∑ C Ai + ∑ C Ei + ∑ (C Pi + C Si )
i =1 i =1 i =1
where CAi is the transmission time of an alarm frame generated
in the ith fire detector, CEi is the transmission time of an
extinction frame generated in the ith actuator, CPi and CSi are
the transmission times of a poll frame and status frame that are
transmitted to a fire detector or an actuator, and Nd and Na are
the number of fire detectors and actuators that are connected to
the CAN-based fire-detection system. For example, suppose
that the data length of all frames is 1 byte, and that 8,000 fire
detectors and 2,000 actuators are connected to the CAN-based
fire-detection system. Here, the data length has been set to 1
byte because that is sufficient to transmit the data measured by
the fire detectors. In this case, CAi is 180 µs from (1), and TP
and TA are 5.4 s from (2). Hence, if TP and TA are greater than
5.4 s, all frames can be transmitted within one cycle and the
performance of the CAN-based fire-detection system can be
assured.
The National Fire Protection Association (NFPA) suggests
that the fire warning response time for a fire alarm signal,
fire and when actuators start to extinguish it, should be less
than 90 s [5][17]. This response time can be regarded as the
maximum allowable delay for a CAN-based fire-detection
system. The transmission delay of 5.4 s in the above example
is much less than the maximum allowable delay, so the
performance of our CAN-based fire-detection system meets
the NFPA guidelines. As far fewer fire detectors and actuators
are installed in home automation systems than in our example,
our CAN-based fire-detection system can be effectively
implemented in such systems.
IV. EXPERIMENTAL EVALUATION OF A CAN-BASED FIRE-
DETECTION SYSTEM
The experimental model shown in Fig. 8 is used to evaluate
(2) the feasibility of our CAN-based fire-detection system. In this
experimental setup, four communication modules for one PC-
1098 IEEE Transactions on Consumer Electronics, Vol. 50, No. 4, NOVEMBER 2004

based receiver, two smoke detectors, one gas detector, and an detector did not receive an extinction frame from the receiver
additional communication module to emulate the actuator are within a period of TA, it repeated its transmission of an alarm
connected to the CAN. The transmission speed of the CAN is frame with a new detection value. Here, TA was set to 10 ms.
set to 500 kbps, and one CAN repeater is used to expand the Fig. 10 shows a transmission frame in our CAN-based fire-
transmission range of the network. Also, a PC-based receiver detection system measured by an oscilloscope. Fig. 10(a)
is programmed using CAPL by CANalyzer. The smoke shows the poll and status frames that are periodically
sensors are NEMOTO NIS-05As, and the gas sensor is a exchanged between the receiver and fire detector. If a fire
NEMOTO NAP-55A. An Atmel AT89C51CC01 occurred, the fire detector transmits an alarm frame to the
microcontroller is used as a communication module that acted receiver, as shown in Fig. 10(b). When the receiver
as a CAN 2.0B controller, and a Philips PCA82C250 is used acknowledges that a fire have occurred after receiving a fire
as the CAN transceiver for signaling. detection value beyond a threshold value, it transmits an
Fig. 9(a) shows the operation algorithm of the PC-based extinction frame to the corresponding fire detectors. Here, the
receiver in the experimental model. As shown in the figure, the response time of the fire alarm signal is 0.27 ms, which is
receiver transmits poll frames to the fire detectors in a given much smaller than the NFPA requirement of 90 s [5][17].
order, once every TP cycles. At this time, if a collision
occurred, IDs are compared, and the frame with the highest
priority ID is transmitted first. The fire level ID of the frame Power on
that lost the competition is increased, and it then waits until the
transmission medium is idle. If an alarm frame with a fire
quantity that exceeds the threshold value is detected during the yes Is alarm frame
carrier sensing, an extinction frame is transmitted to activate detected?
the corresponding actuators. In these tests, TP is set to 10 ms, no
which is sufficient because all frames generated in the
Transmit Transmit
experimental model are transmitted within this time. extinction frame poll frame
Fig. 9(b) shows the operation algorithm for a fire detector
and an actuator in the experimental model. As shown in the
Is arbitration no Increase no Is arbitration
figure, when the fire detector received a poll frame under a no-
won? fire level ID won?
fire condition, it transmitted a status frame containing its
detection value. If the fire detector perceived a fire, it yes yes
transmitted an alarm frame that included the changed detection Is status frame yes
value and waited to receive an extinction frame. If the fire detected?
no
terminator Report Report
NIS-50A fire occurrence error of node
smoke detector
smoke sensor
CAN module
(a) Operation algorithm of the PC-based receiver
actuator
CAN module Power on

T89C51cc01 82C250 yes

CAN Is fire detected?
Repeater NAP-55A
gas detector no

gas sensor Transmit Is poll frame no

alarm frame received?
CAN module
smoke detector yes
smoke sensor
Increase yes Is arbitration Transmit
CAN module fire level ID won? status frame

no

Is extinction frame no
PC based receiver received within TA?
yes

terminator (b) Operation algorithm of the fire detector or actuator

Fig. 8. Experimental testbed of a CAN-based fire-detection system. Fig. 9. Operation algorithm of a CAN-based fire detector.
K. C. Lee and H.-H. Lee: Network-based Fire-Detection System via Controller Area Network for Smart Home Automation
Fig. 11 shows a measured analog fire value from a smoke
sensor obtained by the receiver. As shown in the figure, if the
receiver continuously receives a smoke voltage greater than 4
V from the smoke sensor, it judges that a fire has occurred and
transmits an extinction frame to the corresponding actuators. If
it receives a value lower than 4 V from the smoke sensor, it
judges that the fire has been extinguished.
In a conventional fire-detection system, when the receiver
receives a value greater than 4 V owing to a temporary
malfunction from noise, it decides that a fire has occurred and
activates the actuators. However, A CAN-based fire detector
transmits analog values of fire detection data to the receiver
when it detects a fire, which can be used to avoid false alarms.
Also, a CAN-based receiver receives fire detection data from
several fire detectors, and can use this information to evaluate
whether a fire has indeed occurred. Therefore, owing to its
advanced features, a CAN-based fire-detection system is more
accurate than a conventional fire-detection system. In addition,
if two or three fire detectors are installed at any particular
point, a CAN-based fire-detection system can accurately detect
a fire despite of a malfunction in any one of fire detectors.
V. SUMMARY AND CONCLUSIONS
This paper presents a CAN-based fire-detection system and
verifies that such a system can be used in a smart home. We
describe the structure of a CAN-based fire-detection system
and the design method of the CAN. Also, an experimental
model is used to evaluate the performance of the system.
From several experiments, it can be said that the system
(a) Normal state
(b) Fire detection
Fig. 10. Message transmission in the CAN-based fire-detection system.
1099
sign of fire end of extinction
voltage(V)
If a sensor value is less than 4V,
we define that a fire is occurred
time
Fig. 11. Message transmission for fire detection using an actual
sensor value.
satisfies the NFPA requirements in that the fire warning
response time is very short. Also, as the receiver obtains fire
detection data directly from the fire detectors, it can be said
that the system is more accurate than the conventional fire-
detection systems. In addition, if two or three fire detectors are
installed at any particular point, the CAN-based fire-detection
system can accurately detect a fire despite of a malfunction
occurring in any one of fire detectors.
Especially, the price of a CAN microcontroller, which can
be integrated into one-chip by semiconductor manufacturing
companies, is low, and many CAN development tools are
coming onto the market. Hence, a fire-detection system based
on the CAN has the advantage that the implementation of
sensors or actuators is straightforward as compared to a system
using an Ethernet that is the lower layer of the BACNet. This
makes it easier to design a home automation system.
However, this paper focuses only on the basic structure of a
CAN-based fire-detection system. For practical application of
these research results, the implementation of an application
layer and a power supply should also be studied. In addition,
research into the redundancy of the communication module
will be required to obtain the necessary fault-tolerance
properties required in a CAN-based fire-detection system.
REFERENCE
[1] D. Snoonian, “Smart buildings,” IEEE Spectr., vol. 40, no. 8, pp. 18-
23, Aug. 2003.
[2] R. Nunes, J. Delgado, “An Architecture for a Home Automation
System”, IEEE Intl Conf. Electron. Circuits Syst., pp. 259-262, 1998.
[3] K.S. Lee, K.C. Lee, S. Lee, K.T. Oh, and S.M. Baek, “Network
Configuration Technique for Home Appliances based on LnCP,” IEEE
Trans. Consum. Electron., vol. 49, no. 2, pp. 367-374, 2003.
[4] K. Wacks, “Home Systems Standards: Achievements and Challenges,”
IEEE Comm. Mag., vol. 40, no. 4, pp. 152-159, 2002.
[5] W.S. Song, S.H. Hong, and K.D. Song, “Performance analysis of
BACnet-based fire detection/monitoring system in the high-rise
building”, Journal of the Architectural Institute of Korea, vol. 18, no. 1,
pp. 171-178, 2002.
[6] N. Sriskanthan, F. Tan, and A. Karande, “Bluetooth based home
automation system,” Microprocessor. Microsyst., vol. 26, pp. 281-289,
2002.
1100
[7] D. Snoonian, “Smart Buildings,” IEEE Spectr., vol. 40, no. 8, pp. 18-
23, 2003.
[8] Road Vehicles-Interchange of Digital Information-Controller Area
Network for High-Speed Communication, ISO 11898, 1994.
[9] T. C. Hooi, M. Singh, Y. K. Siah, and A. R. Ahmad, “Building Low-
Cost Intelligent Building Components with Controller Area Network
(CAN) Bus,” Proceedings of IEEE Region 10 International Conference
on Electrical and Electronic Technology, pp. 466-468, 2001.
[10] CiA DSP 416-1 V 1.0 part 1: CANopen application profile for building
door control - General definitions and basic principles, 2003.
[11] G. Cena and A. Valenzano, “A protocol for automatic node discovery in
CANopen network,” IEEE Trans. Ind. Electron., vol. 50, no. 3, pp. 419-
430, 2003.
[12] D.T. Gottuk, M.J. Peatross, R.J. Roby, and C.L. Beyler, “Advanced fire
detection using multi-signature alarm algorithms,” Fire Saf. J., vol. 37,
no. 4, pp. 381-394, 2002.
[13] K. Tindell and A. Burns, Guaranteed Message Latencies for
Distributed Safety-Critical Hard Real Time Control Networks,
Technical Report YCS 94-229, Dept. Computer Science, Univ. of York,
York, UK, 1994.
[14] M. Ellims, S. Parker, and J. Zurlo, “Design and Analysis of a Robust
Real-Time Engine Control Network,”IEEE Micro, vol. 22, no. 4, pp. 20-
27, 2002.
[15] K. Tindell, H. Hansson, and A. Wellings, “Analyzing real-time
communications: controller area network (CAN),” Proceed. IEEE Real-
Time Syst. Sym., pp. 259-263, 1994.
[16] W. Stalling, Local and Metropolitan Area Networks, 6th Ed. Macmillan
Publishing Company, 2000.
[17] NFPA, National Fire Alarm Code, 1996.
IEEE Transactions on Consumer Electronics, Vol. 50, No. 4, NOVEMBER 2004
Kyung Chang Lee (S’00-A’04) was born in Pusan,
Korea, in 1971. He received the B.S., M.S. and Ph.D.
degrees in mechanical engineering from Pusan National
University, Pusan, Korea, in 1996, 1998, and 2003,
respectively. He is a Researh Aassociate in the Network-
based Automation Research Center, University of
Ulsan, Ulsan, Korea. From 1997 to 2003, he was a
research assistant in the Mechanical Engineering and
Technology Research Information Center, Pusan, Korea. His research
interests are industrial network, in-vehicle network, home network, wireless
network, networked control system, network management by intelligent
techniques. He is a member of Institute of Electrical and Electronics
Engineers, Korean Society of Precision Engineers, and Institute of Control,
Automation, and Systems Engineers.
Hong Hee Lee (M’87) was born in Seoul, Korea, in
1957. He received the B.S., M.S., and Ph.D. degrees
from Seoul National University, Seoul, Korea, in 1980,
1982, and 1990, respectively. He is a professor in the
School of Electric-Electronic Information System
Engineering, University of Ulsan, Ulsan, Korea. He is
also Director of Network-based Automation Research
Center (NARC) which is sponsored by Ministry Of
Science and Technology (MOST). His research interests is power electronics,
network-based motor control, and control network. He is a member of
Institute of Electrical and Electronics Engineers, Korean Institute of Power
Electronics, Korean Institute of Electrical Engineers, and Institute of Control,
Automation, and Systems Engineers.
 International Journal of Advanced Science and Technology
Vol. 15, February, 2010

A Review on Security in Smart Home Development

Rosslin John Robles1 and Tai-hoon Kim1

1
School of Multimedia, Hannam University, Daejeon, Korea
rosslin_john@yahoo.com, taihoonn@hannam.ac.kr

Abstract
A smart home or building is a home or building, usually a new one that is equipped with
special structured wiring to enable occupants to remotely control or program an array of
automated home electronic devices by entering a single command. Conventional security
systems keep homeowners, and their property, safe from intruders. A smart home security
system, however, offers many more benefits. Home automation technology notifies
homeowners of any problems, so that they can investigate. In this paper, we discuss smart
home and security, we also review the tool related to smart home security.

Keywords: Smart Home, Security, Automation, Protection

1. Introduction
Smart homes connect all the devices and appliances in your home so they can
communicate with each other and with you. Anything in your home that uses electricity can
be put on the home network and at your command. Whether you give that command by voice,
remote control or computer, the home reacts. Most applications relate to lighting, home
security, home theater and entertainment and thermostat regulation. Security has been an
important issue in the smart home applications. Conventional security systems keep
homeowners, and their property, safe from intruders. A smart home security system, however,
offers many more benefits. On the following chapters, we discuss smart home, smart home
security and related tools in smart home security.

2. Smart Home Systems

A smart home or building is a home or building, usually a new one that is equipped with
special structured wiring to enable occupants to remotely control or program an array of
automated home electronic devices by entering a single command. For example, a
homeowner on vacation can use a Touchtone phone to arm a home security system, control
temperature gauges, switch appliances on or off, control lighting, program a home theater or
entertainment system, and perform many other tasks. [1]
The field of home automation is expanding rapidly as electronic technologies converge.
The home network encompasses communications, entertainment, security, convenience, and
information systems. [1]
A technology known as Powerline Carrier Systems (PCS) is used to send coded signals
along a home's existing electric wiring to programmable switches, or outlets. These signals
convey commands that correspond to "addresses" or locations of specific devices, and that
control how and when those devices operate. A PCS transmitter, for instance, can send a

13
International Journal of Advanced Science and Technology
Vol. 15, February, 2010

signal along a home's wiring, and a receiver plugged into any electric outlet in the home could
receive that signal and operate the appliance to which it is attached. [1]
One common protocol for PCS is known as X10, a signaling technique for remotely
controlling any device plugged into an electrical power line. X10 signals, which involve short
radio frequency (RF) bursts that represent digital information, enable communication between
transmitters and receivers.
In Europe, technology to equip homes with smart devices centers on development of the
European Installation Bus, or Instabus. This embedded control protocol for digital
communication between smart devices consists of a two-wire bus line that is installed along
with normal electrical wiring. The Instabus line links all appliances to a decentralized
communication system and functions like a telephone line over which appliances can be
controlled. The European Installation Bus Association is part of Konnex, an association that
aims to standardize home and building networks in Europe.

Figure 1. Smart Home Technology Automation

Echelon Corp., the creator of the LonWorks system, is helping drive adoption of an open
interoperability standard among vendors in the control networks industry. LonWorks is an
open standard for network automation and control for the building, transportation, industrial
and home markets. The American National Standards Institute (ANSI) has adopted the
protocol underlying LonWorks control networks as an industry standard. The LonMark
Interoperability Association is made up of more than 200 controls companies mission
working on standard to integrate multi-vendor systems based on LonWorks networks. [2]

2.1. Smart Home Software and Technology

Smart home technology was developed in 1975, when a company in Scotland developed
X10. X10 allows compatible products to talk to each other over the already existing electrical

14
 International Journal of Advanced Science and Technology
Vol. 15, February, 2010

wires of a home. All the appliances and devices are receivers, and the means of controlling
the system, such as remote controls or keypads, are transmitters. If you want to turn off a
lamp in another room, the transmitter will issue a message in numerical code that includes the
following:
 An alert to the system that it's issuing a command,
 An identifying unit number for the device that should receive the command and
 A code that contains the actual command, such as "turn off."
All of this is designed to happen in less than a second, but X10 does have some limitations.
Communicating over electrical lines is not always reliable because the lines get "noisy" from
powering other devices. An X10 device could interpret electronic interference as a command
and react, or it might not receive the command at all. While X10 devices are still around,
other technologies have emerged to compete for your home networking dollar.
Instead of going through the power lines, some systems use radio waves to communicate,
which is also how WiFi and cell phone signals operate. However, home automation networks
don't need all the juice of a WiFi network because automation commands are short messages.
The two most prominent radio networks in home automation are ZigBee and Z-Wave. Both
of these technologies are mesh networks, meaning there's more than one way for the message
to get to its destination.

Figure 2. The dots represent devices that could be connected to your smart home
network.

Z-Wave uses a Source Routing Algorithm to determine the fastest route for messages.
Each Z-Wave device is embedded with a code, and when the device is plugged into the
system, the network controller recognizes the code, determines its location and adds it to the
network. When a command comes through, the controller uses the algorithm to determine
how the message should be sent. Because this routing can take up a lot of memory on a

15
International Journal of Advanced Science and Technology
Vol. 15, February, 2010

network, Z-Wave has developed a hierarchy between devices: Some controllers initiate
messages, and some are "slaves," which means they can only carry and respond to messages.
ZigBee's name illustrates the mesh networking concept because messages from the
transmitter zigzag like bees, looking for the best path to the receiver. While Z-Wave uses a
proprietary technology for operating its system, ZigBee's platform is based on the standard set
by the Institute for Electrical and Electronics Engineers (IEEE) for wireless personal
networks. This means any company can build a ZigBee-compatible product without paying
licensing fees for the technology behind it, which may eventually give ZigBee an advantage
in the marketplace. Like Z-Wave, ZigBee has fully functional devices (or those that route the
message) and reduced function devices (or those that don't).
Using a wireless network provides more flexibility for placing devices, but like electrical
lines, they might have interference. Insteon offers a way for your home network to
communicate over both electrical wires and radio waves, making it a dual mesh network. If
the message isn't getting through on one platform, it will try the other. Instead of routing the
message, an Insteon device will broadcast the message, and all devices pick up the message
and broadcast it until the command is performed. The devices act like peers, as opposed to
one serving as an instigator and another as a receptor. This means that the more Insteon
devices that are installed on a network, the stronger the message will be. [3]

2.2. Setting Up a Smart Home

X10, Insteon, ZigBee and Z-Wave just provide the technology for smart home
communication. Manufacturers have made alliances with these systems to create the products
that use the technology. Here are some examples of smart home products and their functions.
 Cameras will track your home's exterior even if it's pitch-black outside.
 Plug your tabletop lamp into a dimmer instead of the wall socket, and you can
brighten and dim at the push of a button.
 A video door phone provides more than a doorbell -- you get a picture of who's at the
door.
 Motion sensors will send an alert when there's motion around your house, and they
can even tell the difference between pets and burglars.
 Door handles can open with scanned fingerprints or a four-digit code, eliminating the
need to fumble for house keys.
 Audio systems distribute the music from your stereo to any room with connected
speakers.
 Channel modulators take any video signal -- from a security camera to your favorite
television station -- and make it viewable on every television in the house.
 Remote controls, keypads and tabletop controllers are the means of activating the
smart home applications. Devices also come with built-in web servers that allow you
to access their information online.
The keypad will send a message to your lamp. These products are available at home
improvement stores, electronics stores, from technicians or o-nline. Before buying, check to
see what technology is associated with the product. Products using the same technology

16
 International Journal of Advanced Science and Technology
Vol. 15, February, 2010

should work together despite different manufacturers, but joining up an X10 and a Z-Wave
product requires a bridging device.
In designing a smart home, you can do as much or as little home automation as you want.
You could begin with a lighting starter kit and add on security devices later. If you want to
start with a bigger system, it's a good idea to design carefully how the home will work,
particularly if rewiring or renovation will be required. In addition, you'll want to place
strategically the nodes of the wireless networks so that they have a good routing range.
The cost of a smart home varies depending on how smart the home is. One builder
estimates that his clients spend between $10,000 and $250,000 for sophisticated systems. If
you build the smart home gradually, starting with a basic lighting system, it might only be a
few hundred dollars. A more sophisticated system will be tens of thousands of dollars, and
elements of home theater systems raise the cost of a system about 50 percent. [3]

2.3. Benefits of Smart Home

Smart homes obviously have the ability to make life easier and more convenient. Home
networking can also provide peace of mind. Whether you're at work or on vacation, the smart
home will alert you to what's going on, and security systems can be built to provide an
immense amount of help in an emergency. For example, not only would a resident be woken
with notification of a fire alarm, the smart home would also unlock doors, dial the fire
department and light the path to safety.
Smart homes also provide some energy efficiency savings. Because systems like Z-Wave
and ZigBee put some devices at a reduced level of functionality, they can go to "sleep" and
wake up when commands are given. Electric bills go down when lights are automatically
turned off when a person leaves the room, and rooms can be heated or cooled based on who's
there at any given moment. One smart homeowner boasted her heating bill was about one-
third less than a same-sized normal home. Some devices can track how much energy each
appliance is using and command it to use less.
Smart home technology promises tremendous benefits for an elderly person living alone.
Smart homes could notify the resident when it was time to take medicine, alert the hospital if
the resident fell and track how much the resident was eating. If the elderly person was a little
forgetful, the smart home would perform tasks such as shutting off the water before a tub
overflowed or turning off the oven if the cook had wandered away. It also allows adult
children who might live elsewhere to participate in the care of their aging parent. Easy-to-
control automated systems would provide similar benefits to those with disabilities or a
limited range of movement.

3. Security Technology
As the technology mature and the interest on the internet increases, network security has
become a major concern to companies throughout the world. The fact that the information
and tools needed to penetrate the security of corporate networks are widely available has
increased that concern. Because of this increased focus on network security, network
administrators often spend more effort protecting their networks than on actual network setup
and administration. Tools that probe for system vulnerabilities, such as the Security
Administrator Tool for Analyzing Networks, and some of the newly available scanning and
intrusion detection packages and appliances, assist in these efforts, but these tools only point

17
International Journal of Advanced Science and Technology
Vol. 15, February, 2010

out areas of weakness and may not provide a means to protect networks from all possible
attacks. Thus, as a network administrator, you must constantly try to keep abreast of the large
number of security issues confronting you in today’s world. [4]

3.1. Protecting Confidential Information

Confidential information can reside in two states on a network. It can reside on physical
storage media such as a hard drive or memory, or it can reside in transit across the physical
network wire in the form of packets. These states present multiple opportunities for attacks
from users on your internal network, as well as those users on the Internet. We are primarily
concerned with the second state, which involves network security issues. The following are
five common methods of attack that present opportunities to compromise the information on
your network: [4]
• Network packet sniffers
• Password attacks
• IP spoofing
• Man-in-the-middle attacks
• Distribution of sensitive internal information to external sources
When protecting your information from these attacks, your concern is to prevent the theft,
destruction, corruption, and introduction of information that can cause irreparable damage to
sensitive and confidential data. This section describes these common methods of attack and
provides examples of how your information can be compromised. [4]

4. Smart Home Security

Conventional security systems keep homeowners, and their property, safe from intruders.
A smart home security system, however, offers many more benefits. Home automation
technology notifies homeowners of any problems, so that they can investigate. Artificial
intelligence programs keep track of the homeowner's habits, and other important information,
and notify emergency personnel when necessary. [5]

4.1. Smart Home Fire Protection

Smart home security system offers much more protection than the typical fire alarm. This
type of system checks carbon monoxide levels as well as watching for signs of fire and
monitors all areas of the home. In the event of a fire, the smart home security system can alert
the homeowner and notify emergency services. Artificial intelligence programs are even able
to pinpoint the location of the fire, and provide that information to fire department personnel
as they respond. [5]

4.2. Access Control

Security codes, motion detectors, and cameras provide information to a smart home
security system, allowing it to determine whether an individual is a resident, a cleared visitor,
or an intruder. Motion detectors trigger an alert, letting the artificial intelligence program
know that there is someone or something to be evaluated. Facial recognition software and

18
 International Journal of Advanced Science and Technology
Vol. 15, February, 2010

security codes allow the security system to allow residents into the home, while based on pre-
programmed information restrict access to other individuals. [5]
Whenever smart home security system detects someone who is unknown, it can provide
video of the visitor to the homeowner. Visitors that are welcome can be given clearance and
allowed in the house remotely. Unwelcome visitors can be ignored, and individuals
attempting to break in will trigger a call to the police. [5]

4.3. Artificial Intelligence Programs Protect the Homeowner

Intruders and fires are not the only dangers in a home. A smart home security system also
protects residents from unanticipated health problems. Using the same cameras and motion
detectors that protect the outside of a home, smart houses can learn about the habits and
normal movements of the residents. When the resident does something unexpected, and does
not resume normal activities, the smart home can alert family members or emergency
services. This aspect of a smart home is particularly helpful for the elderly, or those in fragile
health. [5]

5. Technology and Researches in Smart Home Security

Many smart home devices provide home automation technology, but the smart home
security system offers many benefits that can ensure the safety of the homeowner. In this
section, we review the tools related to Smart Home Security.

5.1. Smart Home Networks User Authentication Using Neural Network

This authentication scheme has two phases: the user registration phase and user
authentication phase. First, authorized users have to register in the authentication system by
giving their username and password. In second phase, i.e., the user authentication phase, the
system validates the legitimacy of the users. [6]
In the user registration phase, the system administrator obtains the training patterns from
usernames and passwords to train the neural network. The registration process is described as
follows:
1. Each user chooses a proper username and password and gives them to the system
administrator.
2. The system applies a one-way hash function to the username and password and the
result is used as the training pattern. So, the training pattern consists of hashed
username, as the input of the neural network, and the corresponding hashed
password, as the desired output of the neural network.
3. Before training the neural network, the system needs to normalize the ASCII codes of
the characters of the training patterns.
4. The system administrator uses these training patterns for training the RBF network.
After training process, the system administrator stores the RBF network weights in
the system.

19
International Journal of Advanced Science and Technology
Vol. 15, February, 2010

In the user authentication phase, the authentication system uses the trained RBF network
and applies the same one-way hash function to authenticate the legality of the users. The
authentication process is described as follows: [6]
1. The system applies the same hash function on the entered username and password.
2. The system extracts an output through the trained neural network.
3. The system compares the output of the RBF network with the hashed password. If the
results are equal, the user is recognized as an authorized user. Otherwise, the user is
rejected as an illegal user.

5.2. Sentry@Home
Smart Home environments typically are equipped with different kinds of sensors and
tracking devices for context-aware service provisioning. While on the one hand, people want
to take advantage of the comfort and added value of personalized context-aware services,
privacy and traceability becomes a serious concern on the other hand. The question arises,
how we can build up trust into inherently untrusted services in a potentially hostile
environment? How can it be guaranteed that eventually all sensitive data is deleted or safely
stored away? The Sentry@HOME concept, as part of our User-centric Privacy Framework,
addresses these concerns.
Sentry@HOME is designed to become an integral part of the user’s home environment;
seamlessly embedded into the Smart Home software infrastructure.. The Smart Home itself
then can be leveraged to act as a privacy proxy for a tracked individual. On behalf of the user
it constitutes the central privacy enforcement point for all privacy-relevant accesses to private
or sensitive data. We are confident that our contribution, the combination of Smart Homes
and a privacy-aware infrastructure, substantially adds to the success of personalized pervasive
computing systems.

5.3. Defending DDoS Attacks Caused by Spam

Distributed Denial of Service (DDoS) attack is a large scale, coordinated attack on the
availability of services at a victim system or network resource. DDOS attack through spam
mail is one of the new versions of common DDoS attack. In this type, the attacker penetrates
the network by a small program attached to the spam mail. After the execution of the attached
file, the mail server resources will be eaten up by mass mails from other machines in the
domain results denial of services.
This defense mechanism is a multi layer approach to defend the DDoS attack caused by
spam mails. This approach was implemented in the mail system and monitored the results.
The result shows that our approach is very effective. The approach has six layers. This
approach is a combination of fine tuning of source filters, content filters, network monitoring
policy, general email policies, educating the user & timely logical solutions of the network
administrator. Fine tuning of source filters reject the incoming connections before the spam
mail delivery. The content filters analyses the contents of the mails and blocks the incoming
unwanted mails. Network monitoring approach provides general solution to identify the
attacks prior to the attack and also during the attack. Business houses should educate the user
about possible attack scenarios & reacting ways to it. The logical solutions of the network
administrator play an important role during the attack period and even post attack period. The

20
 International Journal of Advanced Science and Technology
Vol. 15, February, 2010

combination of these layers provides best methodology to stop the DDoS attacks established
though spam mails.

6. Conclusion
Smart Home is a residence that uses a Home Controller to integrate the residence's various
home automation systems. The most popular Home Controllers are those that are connected
to a Windows based PC during programming only, and are then left to perform the home
control duties on a stand-alone basis. Integrating the home systems allows them to
communicate with one another through the home controller, thereby enabling single button
and voice control of the various home systems simultaneously, in preprogrammed scenarios
or operating modes. Security has been an important issue in the smart home applications. In
this paper, we discussed smart home and security, we also review the tool related to smart
home security.

References

[1] Vendela Redriksson (2005) "What is a Smart Home or Building" http://searchcio-

midmarket.techtarget.com/sDefinition/0,,sid183_gci540859,00.html#
[2] SearchCIO-Midmarket.com Definitions - smart home or building http://searchcio-
midmarket.techtarget.com/sDefinition/0,,sid183_gci540859,00.html
[3] Molly Edmonds “How Smart Homes Work - Setting Up a Smart Home”
http://home.howstuffworks.com/home-improvement/energy-efficiency/smart-home1.htm
[4] CISCO Press, "Internetworking Technologies Handbook", Chapter 51 - Security Technologies
http://users.freenet.am/~file/DownDB/CISCO_PDF/SecurityTechnologies_CISCO.pdf
[5] Victoria Nicks (2009) "AI Enhances the Smart Home Security System"
http://artificialintelligence.suite101.com/article.cfm/ai_enhances_the_smart_home_security_system
[6] Shahbaz Zahr Reyhani and Mehregan Mahdavi "User Authentication Using Neural Network in Smart Home
Networks", International Journal of Smart Home, Vol. 1, No. 2, July, 2007
[7] Susana Alcalde Bagüés, Andreas Zeidler, Fernandez Valdivielso, Ignacio R. Matias, "Sentry@Home -
Leveraging the Smart Home for Privacy in Pervasive Computing", International Journal of Smart Home, Vol.
1, No. 2, July, 2007
[8] Dhinaharan Nagamalai, Cynthia Dhinakaran and Jae Kwang Lee, "Novel Mechanism to Defend DDoS Attacks
Caused by Spam", International Journal of Smart Home, Vol. 1, No. 2, July, 2007

21
International Journal of Advanced Science and Technology
Vol. 15, February, 2010

Authors

Rosslin John Robles

He received his B.S. in Information Technology from Western

Visayas College of Science and Technology, Philippines. He is
currently a Multimedia integrate Masters-Ph.D. Student at
Hannam University, Korea. His research interests are Software
Engineering, Web Development and IT Security.

Tai-hoon Kim

He received B.E., M.E., and Ph.D. degrees from Sungkyunkwan

University. Now he is a professor, School of Information &
Multimedia, Hannam University, Korea. His main research areas
are security engineering for IT products, IT systems, development
processes, and operational environments.

22
Nama Anggota: Stefanus Willy Prasetyo (18/431117/TK/47710)
Ananda Fikri Nugroho (18/424990/TK/46685)
Andikha Wira Perkasa (18/424991/TK/46686)
Iwan Arif Setiaji (18/425004/TK/46699)
Abdul Rohman(18/431083/TK/47676)
Ikhsan Galih R.(18/428655/TK/47157)
Muhammad Rifqi A.(18/431105/TK/47698)
Zido Yuwazama (15/378802/TK/42744)