See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/47796829

Design and Implementation of a Wireless Sensor

Network for Smart Homes

Article · November 2010

DOI: 10.1109/UIC-ATC.2010.16 · Source: arXiv

CITATIONS READS

10 103

6 authors, including:

Feng Xia
Dalian University of Technology
205 PUBLICATIONS 1,434 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Feng Xia
letting you access and read them immediately. Retrieved on: 03 August 2016
 Design and Implementation of a Wireless Sensor Network for Smart Homes

Ming Xu1, Longhua Ma1, Feng Xia2, Tengkai Yuan1, Jixin Qian1, Meng Shao3
1
Department of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China
e-mail: lhma@iipc.zju.edu.cn
2
School of Software, Dalian University of Technology, Dalian 116620, China
e-mail: f.xia@ieee.org
3
Computer Centre, Hangzhou First People’s Hospital, Hangzhou 310006, China

Abstract—Wireless sensor networks (WSNs) have become population, such as electroencephalograms,

indispensable to the realization of smart homes. The objective electrocardiograms, electromyograms and so on, which uses
of this paper is to develop such a WSN that can be used to ZigBee-based WSNs. ZigBee is particularly suited for the
construct smart home systems. The focus is on the design and implementation of a wide range of low cost, low power
implementation of the wireless sensor node and the consumption, reliable control and real-time monitoring
coordinator based on ZigBee technology. A monitoring system applications within the smart home situations. The above-
is built by taking advantage of the GPRS network. To support mentioned four application areas are also closely related to
multi-hop communications, an improved routing algorithm the design of a WSN for smart homes.
based on the Dijkstra algorithm is presented. Preliminary
In the past, research in smart home and in-home
simulations have been conducted to evaluate the performance
of the algorithm.
applications was often limited to ZigBee technology, and
gradually other long-distance network technologies such as
Keywords-ZigBee; wireless sensor network; smart home; PSTN [3] and GSM [7] are adopted. It turns out that the use
routing algroithm of these technologies makes information more accessible.
This would significantly improve people’s living quality.
However, as a traditional wired network, PSTN has some
I. INTRODUCTION problems, such as unsatisfactory security assurance,
Wireless technologies have been developing rapidly in inconvenience and high cost. Therefore we need a new
these years. The obvious advantage of wireless transmission solution. Among other choices, the GPRS technology can
is a significant reduction and simplification in wiring and solve these problems. Thanks to its unprecedented ubiquity,
harness [1]. Many communication technologies, such as GPRS is now available almost anytime and anywhere, for
IrDA, Bluetooth and ZigBee, GSM/GPRS (General Packet anybody (being served). Furthermore, the GPRS network has
Radio Service), etc., have been developed for different a highly secure infrastructure, which makes sure that the
situations. Nowadays, a kind of real time systems in which information sent or received cannot be stolen [7]. Based on
multiple sensors connected simultaneously to one gateway these observations, we propose to develop a new system
unit become necessary, and they are transformed into utilizing ZigBee sensor networks and the GPRS network
wireless sensor networks (WSNs). connecting the ZigBee networks to the application server.
In previous work, much research has been done using Details about the design and implementation of the system
wireless sensor technologies. Literature search indicates that are presented in the following sections.
applications using wireless sensor technologies have already
existed in the following four fields [2]: II. SYSTEM DESIGN
(1) Home automation and remote monitoring of houses. As shown in Fig. 1, a star-mesh hybrid topology is
For example, Liang et al [3] developed a system of wireless usually used in the smart home system. It mainly contains
smart home sensor network based on ZigBee and PSTN the following components:
(Public Switched Telephone Network) technologies. (1) ZigBee network coordinator: This special node takes
(2) Environmental monitoring, including humidity,
the responsibilities of controlling data communications,
temperature and radiation. For instance, Rosiek and Batlles
establishing communication links and protecting
[4] presented a system of data-acquisition from remote
meteorological stations using the mobile communication equipments inside the network.
networks (more specifically, GPRS). (2) ZigBee node: A node is mainly composed of various
(3) Fault tracking and fault management. For example, in sensors and a ZigBee wireless module. In practice, nodes
[5], the authors developed an online diagnosis and real time can be deployed to establish a network with a star-like,
warning system for vehicles using 3G technologies and mesh, or hybrid topological structure. In the monitoring area,
GPRS communications. ZigBee nodes are scattered according to the distance and all
(4) Health monitoring. For instance, Monton et al [6] sensor data can be sent to the network coordinator though
designed an e-health approach to monitoring data of specific the network.
 (3) GPRS network: Data generated within the ZigBee The three modules can connect and communicate with each
network is transferred to the monitoring center via the other through serial ports.
GPRS network and the Internet.
(4) Monitoring center: A computer/server in the PSTN ZigBee
monitoring center is used to manage the data generated by circuits (g) circuits (b)
all Zigbee networks.
MSP430
ZigBee network GPRS Peripheral AD
node circuits (c) circuits circuits (e)
(a)
node
node GPRS
SIM card DI and DO
coordinator circuits (d) circuits (f)

node Figure 2. Internal structure of coordinator

node
Monitoring As shown in Fig. 2, parts (a), (b) and (e) constitute a
node center typical ZigBee node. Adding parts (c) and (d) to the node, it
becomes a ZigBee coordinator. We add parts (f) and part (g)
Figure 1. Network topology of smart home systems for additional functions. Part (f) can control some target
with 2A current relay and monitor the switch mode of the
In our daily applications, several ZigBee networks target. Part (g) can generate and decode PSTN signals.
would overlap each other and the nodes of the network may Home secure alarm mainframe can be connected to the tip
breakdown [8, 9]. Therefore characteristic identification of and ring port. When the mainframe alarms, part (g) can
the node should be considered in the design to distinguish extract alarm information of Contract ID (CID, one popular
the networks. Redundancy nodes are needed to deploy and alarm protocol) code quickly, and the coordinator will send
the mesh topological structure should be optimized. What’s an alarm message to the monitoring center.
more, the routing-algorithm should be examined to ensure The major function of the coordinator in remote
the communications among nodes. communications is as follows. The GPRS communication
A. Hardware Design unit connects to the MCU through the RS232 connector, and
is responsible for data transmission between the node and
The network node and the coordinator are key
the monitoring server. The data sent by the server will get
components of the system. In this work, we assume that no
matter where the user is, the coordinator will always be into the GPRS communication module by antenna. Useful
connected to the monitoring center/server via a computer that data will be obtained through analyzing the TCP/IP
can access the Internet or the GPRS network. It can obtain all agreement. Response data of the MCU will be modulated to
messages exchanged between the server and the network. GSM signal by the GPRS module and be sent to the server
When the server sends out a command, the CPU of the via the internet using the TCP/IP protocol.
network coordinator will read the content of the command
and get the details by analyzing it, such as turning on the air
conditioner or refrigeration. The main control program
within the network coordinator writes the details to the
ZigBee module through serial ports. Then the ZigBee
module will be responsible for sending the messages to the
family network.
From Fig. 1 we can see that development of the network
coordinator and the ZigBee node is the most important task
for hardware system design. These two components are
basically identical. The only difference is that the latter has
the function of GPRS communications while the former
does not. Therefore, we will focus on describing the design
of the ZigBee network coordinator in this paper. According
to the above description, we can find that the
microcontroller (MCU), the ZigBee module, and the GPRS
module are the most important parts of the network
coordinator. The hardware construction of our coordinator
node adopts MSP430F149 as MCU (from TI), ETRX2 Figure 3. ZigBee circuit diagram
ZigBee module, and MC52i GPRS module (from Siemens).
 Figure 4. GPRS circuit diagram

management server serves to manage the interaction between

We have designed peripheral circuits according to the various modules and databases, to establish databases, and to
functional requirements, and developed a network connect to other modules. The system management server
coordinator by integrating the ZigBee coordinator node and interacts with the users. The system management server
the GPRS module together on a PCB board. Given in Fig. 3 provides a human-machine interface, through which users
and Fig. 4 are some of the circuit diagrams designed for the can configure system parameters, monitor real-time data,
system. Fig. 5 shows the hardware board of the coordinator inquire about historical records, etc. The framework of the
node we developed. software is illustrated in Fig. 6.

System management server

System Operation Query

Database management server

Database server Databases

Socket

Figure 5. Coordinator node developed. Figure 6. Structure of host station software

2) Node Software
B. Software Design The network (coordinator) node software realizes the
1) Monitoring Software collection and transmission of data. Fig. 7 shows the block
The monitoring software on host station (i.e. monitoring diagram of the software. Fig. 7(a) is the main procedure
server) adopts C/S architecture based on Socket flowchart and Fig. 7(b) is the interruption procedure
communication mechanisms of TCP/IP protocol. It is written flowchart. As for sensor nodes, the program realizes
by C# language, using ACCESS database. functions such as data sampling, A/D calculation, I/O
Built on .Net software platform, this software features control, timed sending, and timed hibernating. As for router
independence of platforms and excellent expandability. The nodes and coordinator nodes, it mainly realizes the function
whole monitoring system of the host station mainly consists of data forwarding and path routing. The router nodes and
of two parts: database management server and system the coordinator nodes have the capacity of collecting sensor
management server. The database management server
data. Therefore they can also be treated as sensor nodes.
includes database server and databases. The database
During the interaction between coordinator nodes and the table of distances between nodes by using the following
GPRS network, both of them should follow the same algorithm [11]:
datagram protocol in order to enable the host station to (1) A node is arbitrarily selected as the root node. After
analyze the message more easily. initialized, it will send messages to surrounded nodes asking
for their IDs and location information.
(2) In response to Step (1), the remaining n-1 nodes send
Initialization their IDs and location information to the root node.
(3) The distance table is created after the root node has
received all the information of remaining n-1 nodes.
N As an example, Fig. 8 gives a simple wireless network.
GPRS is connected? Connect GPRS The corresponding distance table is shown in Table I.
Y
N 1
Send GPRS message? Send GPRS message 5 4
2
2
5 2
Y 4 3
N
3 4 5 5
Send ZigBee message? Send ZigBee message 4
5 3
7
4 6
Y 7
2
4 2
Delay 100ms 5
8 4
9
4
(a) 4 10
0
Interruptions in Figure 8. An example network of 10 nodes

N TABLE I. DISTANCES BETWEEN NODES

Receive SPTN message? Connect GPRS
Node ID 1 2 3 4 5 6 7 8 9 10
Y 1 0 4 2 5 5 7 6 8 9 9
2 4 0 2 7 6 5 5 7 7 8
N 3 2 2 0 5 4 5 4 6 7 8
Receive GPRS message? Receive GPRS message 5 7 5 0 3 8 7 7 9 8
4
5 5 6 4 3 0 6 5 2 7 5
Y 6 7 5 5 8 6 0 3 8 2 5
N 7 6 5 4 7 4 3 0 5 4 4
Receive ZigBee message? Receive ZigBee message 8 8 7 6 7 2 8 5 0 6 4
9 9 7 7 9 7 2 4 6 0 4
Y 10 9 8 8 8 5 5 4 4 4 0

Interruptions out After the table of distances is created, we use an

improved Dijkstra algorithm to deduce the optimal path.
(b)
Some symbols and notations used in the algorithm are listed
Figure 7. Flowchart of node software: (a) main procedure; (b) in Table II.
interruption procedure.
TABLE II. SYMBOL DEFINITION
III. ROUTING ALGORITHM Symbol Description
n Number of network nodes
The smart home system using the WSN developed above v Sending node
can be modeled as a (wireless) network, and the routing w Receiving node
point is the node in the network. The traditional Dijkstra k Transmission radius
algorithm [10] generates the shortest path according to the u Relay node
order of increasing path length, and greedily searches path s[i] Visit mark (if not visited, set s[i]=0; else set
s[i]=1)
based on the edges connected with nodes. However, there is
cost[i][j] Distance between node i and node j
no edge in the wireless network. Therefore we propose an dist Distance corresponding to the optimal path from
improved Dijkstra algorithm for the WSN, which obtains the v to w
shortest path in the network.
Assume that there are n nodes in a wireless network, and The improved Dijkstra algorithm is described in the
the location of each node is available. Then we can get the following:
 (1) Initialization: num=0, dist=+∞, s[i] =0, (i=0,1,…n); nodes 3, 5 and 7 consume more energy. This observation
(2) If cost[v][i] ≤k, then set s[i]=1, (i=0,1,…n); could be helpful when placing nodes in a smart home.
(3) If s[w] =1, then dist=cost[v][w]; otherwise go to step

(4) ∀ i ∈ {s[i]=1}, u[num]=1, set s[j]=1 when

(4); IV. CONCLUSIONS

j ∈ {cost[u[num]][j] ≤k}, (j=0,1,…n);

This paper focused on development of the wireless

(5) If s[w] ≠ 1, then num++, repeat step (4); otherwise

sensor node and the coordinator for smart home systems
based on ZigBee technology. Both hardware design and

dist[num] = cos t[v][u[0]]

record the path that fulfills s[w]=1, and software design have been described in detail. A monitoring

∑ cos t[u[ j ]][u[ j + 1]] + cos t[u[num]][w];

system is also built using the GPRS network. To address the
num −1
+
problem of routing for multi-hop communications in smart
home wireless sensor networks, an improved algorithm
j =0 based on the Dijkstra algorithm has been presented. The
(6) For num=num+1, repeat steps (4) and (5) until performance of the algorithm has been evaluated through
all the paths that fulfill s[w]=1 are obtained; preliminary simulations. Our future work is to test and apply
(7) Set dist=min{dist[i], i=0,1,…,num}, and output the the whole system in practice.
corresponding path.
We implemented the above algorithm and conducted ACKNOWLEDGMENT
simulations of a network as given in Fig. 8. Setting k=5, the This work is supported in part by Natural Science
optimal path from sending node 1 to receiving node 10 is to Foundation of China under Grants No. 60474064 and No.
pass a relay node 5, which is better than the path of node 1 60903153, Zhejiang Provincial Natural Science Foundation
→ node 3 → node 7 → node 10, and reduces the number of of China under Grant No.Y107476 and No.R1090052 and
node hops. Therefore, the improved Dijkstra algorithm can the Fundamental Research Funds for the Central Universities.
solve the problem of optimal path selection in a wireless
network, thus providing a feasible routing solution for smart REFERENCES
home systems. [1] W. Ning, Q.Z. Nai, H.W. Mao, “Wireless sensors in agriculture and
To evaluate the performance of routing algorithms, a food industry - Recent development and future perspective,”
general concern would be the energy consumption of the Computers and Electronics in Agriculture, vol. 50, 2006, pp. 1–14.
network nodes. If a node is put into sleep as long as it has no [2] A.R. Al-Ali, M. Al-Rousan, T. Ozkul, “Implementation of
data to receive or send (and all state switching overheads are experimental communication protocol,” Computer Standards &
Interfaces, vol. 28, 2006, pp. 523–530.
negligible), its energy consumption will be approximately
proportional to the number of times it is visited (i.e. the [3] L.L. Liang, L.F. Huang, X.Y. Jiang, V. Yao, “Design and
implementation of wireless smart-home sensor network based on
number of packets it has transferred). In this case, it is ZigBee protocol,” International Conference on Communications,
possible to examine the relative energy consumption of all Circuits and Systems, 2008, pp. 434-438.
nodes by observing the number of times each node is visited. [4] S., Rosiek, F.J. Batlles, “A microcontroller-based data-acquisition
system for meteorological station monitoring,” Energy Conversion
and Management, vol. 49, 2008, pp. 3746–3754.
[5] C.E. Lin, C.C. Li, S.H. Yang, S.H. Lin, C.Y. Lin, “Development of
On-Line Diagnostics and Real Time Early Warning System for
Vehicles,” SICON05 - Sensors for Industry Conference, Houston,
Texas, USA, February 2005, pp. 8-10.
[6] E. Monton, J.F. Hernandez, J.M. Blasco, T. Herve, J. Micallef, I.
Grech, A. Brincat, V. Traver, “Body area network for wireless patient
monitoring,” IET Communications, vol. 2, 2008, pp. 215–222.
[7] B.Z. Yan, H.Y. Zhao, “A Low Cost GSM/GPRS Based Wireless
Home Security System,” IEEE Transactions on Consumer Electronics,
May 2008, pp. 567-572.
[8] L.N. Zhang, X.C. Hong, “The successful application of ZigBee
technology in the smart home,” Science & Technology Information,
vol. 16, 2008, pp. 19-20. (in Chinese)
[9] J.M. Bao, “Research on the Application of ZigBee Technology in
Monitoring and Control System of Greenhouse,” Journal of
Figure 9. Number of times each node is visited. Agricultural Mechanization Research, vol. 2, 2008, pp. 184-187.
[10] C.P. Liu, “Mathematical Modeling,” High Education Press, 2002.
In simulations the sending node and the receiving node [11] R. Yang, T. Yang, “Wireless sensor network based on ZigBee with
are randomly generated. Fig. 9 gives the results with data applications for smart building,” Chinese Agricultural Mechanization,
transformation of 1000 and 3000 times. Noticing that nodes vol. 5, 2005, pp. 76-79. (in Chinese)
3, 5 and 7 have more visited times, we can conclude that