Spread Spectrum Based Energy Efficient Wireless Sensor Networks

ACEEE International Journal on Network Security, Vol 1, No. 2, July 2010

Spread Spectrum Based Energy Efficient
Wireless Sensor Networks
Wg Cdr Varghese Thattil (Retd)1, Dr. N Vasantha2, T. Esther Rani3
1
CVR College of Engineering Electronics and Communication, Hyderabad, India
Email:mailthattil@gmail.com
2
Vasavi College of Engineering, Hyderabad, India
Email: vasanthavasantha@rediffmail.com
3
CVR College of Engineering Electronics and Communication, Hyderabad, India
Email: esterlawrence@rediffmail.com

Abstract—The Wireless Sensor Networks (WSN) is energy saving is one of the key issues in the design of
considered to be one of the most promising emerging systems based on WSNs.
technologies. However one of the main constraints which It has been experimented and seen that the data
is holding back its wide range of applications is the communication of sensor nodes consumes more energy
battery life of the sensor node and thus effecting the
network life. A new approach to this problem has been
than the data processing and therefore there is a need to
presented in this paper. The proposed method is suitable reduce the communication while achieving the desired
for event driven applications where the event occurrence network operation [1]. The energy cost of transmitting
is very rare. The system uses spread spectrum as a means one single bit of information is approximately same as
of communication. that needed for processing a thousand operations in a
typical sensor node [2]. However, since the sensor
Index Terms—Rare event, Wake-up receiver, Single hop, network is very dense; composed of nodes with low
Spread spectrum duty-cycles the medium access decision is a hard
problem. Therefore it is important to know the peculiar
I. INTRODUCTION features of sensor networks including reasons of
Wireless Sensor Networks (WSNs) is considered as potential energy wastes at medium access
a new technological vision because of the rapid communication.
development in miniaturization; low power wireless A. Reasons of Energy Wastage
communication, microsensor, and microprocessor
hardware. WSNs may, in the near future, will provide Energy wastage occurs when there are collided
“Ambient Intelligence” where many different devices packets, which occurs when a receiver node receives
will gather and process information from many more than one packet at the same time or when they are
different sources to both control physical processes and partially overlapped. All packets that cause the
to interact with human users. Potential applications of collision have to be discarded and the re-transmissions
WSNs include environmental monitoring, industrial of these packets are required which increase the energy
control, battlefield surveillance and reconnaissance, consumption. Although some packets could be
home automation and security, health monitoring, and recovered by a capture effect, a number of
asset tracking. requirements have to be met to achieve a successful
Improvements in hardware technology have resulted capturing. The second reason of energy waste is
in low-cost tiny sensor nodes which are composed of overhearing, meaning that a node receives packets that
three basic components: a sensing subsystem for data are destined to other nodes. The third energy waste
acquisition from the surrounding environment, a occurs as a result of control packet overhead. Minimal
processing subsystem for local data processing and number of control packets should be used to make a
storage and a wireless communication sub system for data transmission. One of the major sources of energy
data transmission to a sink node or access point. In waste is idle listening, i.e., listening to an idle channel
addition to this; the energy required to carry out the to receive possible traffic. The last reason for energy
programmed task is supplied from battery pack with a waste is overemitting, which is caused by the
very limited energy budget. In most of the applications transmission of a message when the destination node is
it has been observed that the energy source can not be not ready. Given the above facts, an efficiently
replaced due to hostile environment or other practical designed MAC protocol should reduce these energy
difficulties. However the sensor networks should be wastes to a large extend.
able to sustain till such time the indented task is B. Communication Patterns
completed. Though energy scavenging from external Kulkarni et al. defines three types of communication
environment can be a possible solution, the technology patterns in wireless sensor networks [3]: broadcast,
is not yet developed into a reliable source. Therefore convergecast, and local gossip. Broadcast type of

1

© 2010 ACEEE
DOI: 01.ijns.01.02.01


communication pattern is generally used by a base Additional components like as location finding system
station (sink) to transmit some information to all sensor to determine their position, a mobilizer to change their
nodes of the network. Broadcasted information may location or configuration etc. may be added depending
include query-processing information, program updates upon the applications where they are used.
for sensor nodes, control packets for the whole system Out of the above subsystems, the radio component
etc. In some scenarios, the sensors that detect an account for maximum power consumption. Therefore
intruder communicate with each other locally. This many protocols have been designed with the idea of
kind of communication pattern is called local gossip, minimizing the radio communication this in most of the
where a sensor sends a message to its neighboring cases has been achieved by increased data processing at
nodes within a range. The sensors that detect the the nodes.
intruder, then, need to send what they perceive to the
information center. That communication pattern is
called convergecast, where a group of sensors
communicate to a specific sensor. The destination node
could be a clusterhead, data fusion center, base station.
C. Properties of a Well-defined MAC Protocol
To design a good MAC protocol for the wireless
sensor networks, the following attributes must be
considered [4]. The first attribute is the energy
efficiency. We have to define energy efficient protocols
in order to prolong the network lifetime. Other
important attributes are scalability and adaptability to Figure 1 Wireless Sensor Node Architecture
changes. Some of the reasons behind these network
property changes are limited node lifetime, addition of The communication distances in WSNs are typically
new nodes to the network and varying interference small; a few meters in most of the applications and it
which may alter the connectivity and hence the has been found that power consumed for reception is
network topology. almost same or even more as compared to transmission.
Other typical important attributes such as latency, One of the most power efficient designs for WSN could
throughput and bandwidth utilization may be secondary be the one where in the radio, which consumes
in sensor networks. Contrary to other wireless maximum power could be kept in sleep mode. The
networks, fairness among sensor nodes is not usually a radio should be switched on only when there is a
design goal, since all sensor nodes share a common specific need to transmit or receive data. Therefore the
task. radio should alternate between active and sleep mode
There are mainly two approaches to energy keeping the duty cycle as low as possible.
conservation: In-network processing and through duty However since the WSN is a cooperative system the
cycle controlling. In–network processing is a method sleep/wakeup schedule normally follows a distributive
wherein the amount of data being transmitted is being algorithm. An efficient distributive algorithm should
reduced; either by compression or aggregation ensure that the neighboring nodes in the radio range
techniques. It typically exploits the temporal or spatial should be able to receive whenever a node is
correlation among data acquired by sensor nodes. On transmitting keeping the duty cycle as low as possible.
the other hand, duty cycling schemes define Many WSN applications are event based wherein a
coordinated sleep/wakeup schedules among nodes in communication of data is to be essentially carried out
the network. only on detection of an event. Massively deployed
sensor nodes simultaneously detect the event and hence
II. ENERGY CONSUMPTION REDUCTION can cause duplication of information. Therefore
application layer energy efficiency can be achieved by
A. Possible Energy Saving Approaches in-network processing or data aggregation. The energy
To tackle the problem of energy consumption it is efficiency processing can be either through data quality
essential to know how much power each node or avoiding unwanted data redundancy [7].
component dissipates during normal operating It is important to note at this stage that while
conditions, i.e., which are the power dissipation improving the energy efficiency in a single layer it
characteristics of sensor noses [6]. In a typical wireless should not effect the overall energy consumption in the
sensor node there are four main components: Sensing node as a whole. In many applications, a cross layer
subsystem having sensors for data acquisition and approach may be more suitable to reduce the overall
analog to digital converter, processing subsystem energy consumption in a particular node. The topology
consisting of micro-controller and memory for local control [8] and power management are the two
processing, radio subsystem for wireless data approaches to energy efficiency. In the case of
communication and a power supply unit Figure 1. topology control, the network redundancy is exploited

2
© 2010 ACEEE
DOI: 01.ijns.01.02.01


to improve the energy efficiency whereas in power power management service based on wake-up/sleep
management sleep/wakeup timings are controlled to scheduling.
achieve minimum duty cycle. The study of the energy consumption throughout the
lifespan of a typical surveillance systems network node
B. Power management
shows that most of the energy is used up in operations
Power management schemes need to control when a that do not actively fulfill the system’s purpose. A node
network node should enter a high-power wakeup mode without power management is always turned on, but
and when to enter a low-power sleep mode. Under there is no target most of the time. Hence, most of its
certain conditions, the nodes’ hardware components energy is dissipated in a waiting status. Only one per
may be shut down to low-power state. The low-power cent of the energy is used in actually tracking targets,
to high-power transition is, however, a tricky problem the other 99% of the energy is used in waiting for
because the network node has its CPU halted and is targets to show up. With a rotation based power
unaware of the external events. Power management management, the energy efficiency is much better.
techniques can either be done strictly at MAC layer or
using sleep/wakeup protocols at higher layers (Network D. Low Energy Wakeup Receiver
or application layer) which are integrated on top of L. Gu et al. proposed to use passive components to
MAC layer. In the case of MAC layer power collect radio energy similar to the case of RFID (Radio
management, lower duty cycle is achieved by TDMA frequency Identification) technology. The
(Time Division Multiple Access), Contention-based microcontroller is interrupted when the induced RF
procedures or their combinations. In the case of power input is large enough and this in turn wakes up
TDMA, the nodes need to turn on their radio only the sensor node for data communication. However the
during the allotted time slots. However this poses disadvantage in such a system is that any strong RF
problem of flexibility and scalability. The scheme also signal can wakeup the node and also the range of
has to ensure tight synchronization [9]. operation will be limited. A modified system which
Energy efficiency in the case of Contention-based was proposed is to have an addressing mechanism with
protocols is achieved by tightly integrating channel the help different RF triggering frequencies. The
access functionalities with sleep/wakeup schemes [10]. system will however require additional wakeup
hardware for each frequency [13]. In this case
C. Sleep/wakeup protocols
minimum required active components are added to the
The power management schemes can be broadly passive wake up radio so that messages could be sent to
divided into three categories, depending upon the particular nodes to wake up the nodes. The wakeup
method used to wake up the node. They are on- receiver consists of an impedance matching circuit
demand, schedule based and asynchronous. In the case followed by a voltage multiplier (charge pump). The
of on-demand the node wakes only when there is an charge pump consisting of capacitances along with
external interrupt from another node [11]. The main schottky diodes, increases the weak input voltage to a
problem associated with on-demand schemes is how to level at which the microcontroller can be triggered. The
inform the sleeping node that some other node is schematic diagram has been shown in Figure 2.
willing to communicate with it. To achieve this, Wakeup signal transmitter sends amplitude
additional wake up radios with low data rate and low modulated carrier signal. The received output is finally
power consumption is used for signaling. In the case of passed through a digital comparator to recover the
schedule based approach each node should wake up at envelope. The comparator also helps to have a noise
the same time as its neighbors. However this requires rejection as well as to have square pulse output. In this
rigorous synchronization between the nodes causing architecture digital comparator is the only active
increase in latency. component which will draw power from the source.
The node synchronization between nodes can be Lower distance of wake up radio communication is a
avoided by asynchronous power management. In this big disadvantage in this case.
case, the nodes can wake up when ever they want and
still be able to communicate with their neighbors. This
is achieved by having a pattern of wake up wherein at
least two nodes have an overlapping wake up duration.
In this approach the nodes are expected to wake up
more often and hence the duty cycle tends to be higher
as compared to synchronous protocols [12].
The implementation of the wakeup/sleep scheduling
often involves a timer that wakes up the CPU via an
interrupt. The wake-up/sleep scheduling approach has
some disadvantages. First, the design of a good wake- Figure 2. Wakeup Receiver Block Diagram
up/sleep schedule is often application dependent and
complicated. Hence, it is hard to design a general

3
© 2010 ACEEE
DOI: 01.ijns.01.02.01


III. PROPOSED SCHEME

A. Proposed Scheme Motivation
There are many applications like habitat monitoring
or movement detection etc. of sensor networks where
event monitoring has to be carried out in a large area.
There are applications where the area of interest has to
be continuously monitored for long duration. This
becomes very difficult unless the system is made
automatic. The system should be able to detect the
Figure 4. Sector scanning timings
event and automatically initiate the necessary actions as
required. These can be setting off alarms, send A detailed timing diagram for the system
messages switch on cameras/lights in the required implementation is shown in Figure 5. The antenna
region etc. The advantage of the specific application dwelling time in each sector is 10 seconds. The
can be made use of by having a cross layer protocol so positioning of the antenna in the next sector takes 1
that the power energy efficiency can be maximized. second, however this timing will be lesser if electronic
B. Proposed Scheme Description scanning is used. Once the antenna has been placed to
The proposed scheme is to deploy the sensor nodes the sector; the BS transmitter emits 5 microseconds (µ
in the entire region of interest. Each sensor node will sec) pulses at an interval of 10 milliseconds (msec) for
have its own unique spread spectrum code to identify duration of 1 second. Immediately after that the BS
it. The distance between the nodes have to be decided transmitter shuts off and the receiver is tuned to receive
on the basis of the sensing range of the associated the sensor nodes transmissions.
sensors. The Base station (BS) is having four directive C. Wake-up Receiver
antennas covering the entire region around it. The All sensor nodes will have wake up receivers which
sensor nodes have low power wake-up radios which are tuned to receive the trigger pulses send from the
can be triggered with the help of Base station directive Base Station (BS). As the pulses received at the input
antennas. The deployment plan has been shown in of the wakeup receiver increases the voltage at the
Figure 3 The Base station will sequentially scan the input of the voltage multiplier starts rising; refer Figure
four sectors by sending trigger pulses in the form of 2. When the voltage at the output of voltage multiplier
series of square pulses for few milliseconds and crosses threshold level, the wake up receiver sends an
subsequently the Base Station receivers are switched on interrupt to the microcontroller to initiate the
to receive the transmissions from the sensor nodes. transmission of data signals to BS.
A timing diagram has been shown in Figure.4. The
sensor devices in the sensor nodes will continuously
observe the physical parameters within its range. The
nodes will respond to the wake-up call only incase the
observed parameters crosses a minimum threshold
level. This avoids unwanted transmissions and thus
improving the battery life. At the Base station, the
correlation receivers acquire and track the incoming
signals from the sensor nodes and recover the data.

Figure 5. Base Station and node transmit and receive timings

D. Sensor Node Architecture
The proposed architecture of the sensor node has
been shown in the figure 6. The observed analog
values from the sensor devices are sequentially taken,
converted to digital and stored in the digital storage
registers as obs_value. The obs_value from each sensor
device is stored till such time it is forwarded to digital
comparator. Max_value registers (MVR) shown in the
Figure 3. Deployment of Sensor nodes along with Base architecture are used to store the initial minimum
Station threshold values. These values are given as input to the
digital comparator to compare the obs_value stored in

4
© 2010 ACEEE
DOI: 01.ijns.01.02.01


the corresponding registers of the digital storage. incoming signal with the known PN sequences of the
Incase the obs_value is greater than the Max_value, the sensors and recover the data. Since there is no
MVR is appended with the current obs_value stored in synchronization between the nodes and the Base station
digital storage. The obs_value from each sensor is Asynchronous CDMA communication is required.
converted to digital at an interval of 800 msecs. The
digital comparison is carried out at every 200 msecs. F. Simulation Results
The total observation time per sensor is 44 seconds Sensor node architecture has been designed and
before it being transmitted to BS; Figure 5. implemented using VHDL language in Xilinx 9.2i. The
The hostile environment described is such that the simulation has been carried out by ISE simulator and
intrusion detection is expected to be a rare event. synthesized by using XST. The digital comparison is
Therefore, MVR will not be appended during most of simulated with a timing of 10 nano seconds (equivalent
the sensor observation cycles. The sensor architecture to a real time periodicity of 200 milli seconds of sensor
ensures that the transmission from the sensor node will multiplexing) for each sensor and sweep timing of 800
take place only if one of the sensors obs_value crosses nano seconds has been taken. The timing diagram
the minimum threshold value set in MVR. Minimum shows that the MVR is getting updated only when the
threshold value is selected such that noise and other obs_value crosses the threshold value set in the MVR.
interferences will not trigger unwanted transmissions The wake up trigger required for activating the
from the nodes. This is achieved in the architecture by transmission section will be available only once during
simply setting a flag when the MVR is appended. The the antenna scan of 44 seconds. On arrival of the wake
output from the MVR is taken only if the flag for the up trigger the processor will check whether any of the
corresponding sensor is set. MVRs has been updated. The system will transmit only
once during the antenna provided MVR is updated. The
E. Spread Spectrum Communication
system simulation shows substantial saving in energy
Once the wake up signal has been received, the consumption as the MVR updation takes place rarely.
Max_value stored in MVR is EXORed with the stored Some random values have been given as input from the
PN sequence. The output of the EXOR is BPSK four sensors. The schematic diagram of the sensor node
modulated with the carrier frequency of 10 GHz. This architecture is shown in Figure 6. The selection timing
is further amplified and transmitted to the base station. of obs_value as it crosses the threshold stored MVR is
At the Base station each sensor node is to have a shown in Figure 7.
separate CDMA receiver. These receivers correlate the

Figure 6. Sensor node architecture

Figure 7. Simulation timings at sensor node

5
© 2010 ACEEE
DOI: 01.ijns.01.02.01


G. Energy Saving At The Nodes [4] C. C. Enz, A. El-Hoiydi, J-D. Decotignie, V. Peiris,
“WiseNET: An Ultralow-Power Wireless Sensor Network
In most of the WSN applications the detection of Solution”, IEEE Computer, Volume: 37, Issue: 8, August
an event by a sensor device is a rare event and 2004.
therefore the data need to be transmitted only if an [5] V. Rajendran, K. Obraczka, J.J. Garcia-Luna-Aceves,
event is detected. In the proposed scheme, the sensor “Energy-Efficient, Collision-Free Medium Access Control
for Wireless Sensor Networks”, Proc. ACM SenSys 03,
devices continue to observe the physical parameters Pages:181 - 192, Los Angeles, California, 5-7 November
till such time the wake up signal is arrived. The 2003.
Max_value stored in MVR is appended only when [6] L. Bao and J.J. Garcia-Luna-Aceves, “A New Approach To
obs_value exceeds the minimum threshold value Channel Access Scheduling For Ad Hoc Networks”,
stored in it; which means that only when an event Seventh Annual International Conference on Mobile
occurs. Once the wake up signal is received, the Computing and Networking, pp. 210–221, 2001.
system verifies whether the data stored in the [7] K. Jamieson, H. Balakrishnan, and Y. C. Tay, “Sift: A
corresponding registers of MVR have been MAC Protocol for Event-Driven Wireless Sensor
Networks,” MIT Laboratory for Computer Science, Tech.
appended. The appended data is only selected for
Rep. 894, May 2003
further transmission. This ensures that the data [8] Y.C. Tay, K.Jamieson, H. Balakrishnan, “Collision-
communication from the sensor nodes will take place minimizing CSMA and Its Applications to Wireless Sensor
only if an event occurs. Networks”, IEEE Journal on Selected Areas in
However, the very low level of communication Communications, Volume: 22, Issue: 6, Pages: 1048 –
between the Base station and sensor node can pose a 1057, Aug. 2004.
[9] G. Lu, B. Krishnamachari, C.S. Raghavendra, “An adaptive
problem of identifying dead nodes (dead battery)
energy efficient and low-latency MAC for data gathering in
over a period of time. The system will have a wireless sensor networks”, Proceedings of 18th
mandatory transmission after every ten wakeup International Parallel and Distributed Processing
signal cycles even if there is no event detection. Symposium, Pages: 224, 26-30 April 2004.
[10] T.V. Dam and K. Langendoen, “An Adaptive Energy-
CONCLUSIONS Efficient MAC Protocol for Wireless Sensor Networks”,
The First ACM Conference on Embedded Networked
One of the prime criteria that differentiate WSN Sensor Systems (Sensys‘03), Los Angeles, CA, USA,
nodes to other wireless nodes is the energy November, 2003.
consumption since most of the WSN nodes are [11] P. Lin, C. Qiao, and X. Wang, “Medium access control
suppose to have minimum human interaction which with a dynamic duty cycle for sensor networks”, IEEE
is essential for recharge/replacement of battery. A Wireless Communications and Networking Conference,
method has been suggested which can be very Volume: 3, Pages: 1534 - 1539, 21-25 March 2004.
[12] A. Safwat, H. Hassanein, H. Mouftah, “ECPS and E2LA:
effectively implemented in many event detection
new paradigms for energy efficiency in wireless ad hoc and
applications. The full details of the system has not sensor networks”, IEEE Global Telecommunications
been described, however the paper gives sufficient Conference, GLOBECOM’03, Volume: 6, Pages: 3547 -
scope for the development of the system. 3552, 1-5 December 2003.
[13] S. Cui, R. Madan, A. J. Goldsmith, and S. Lall, “Joint
REFERENCES Routing, MAC, and Link Layer Optimization in Sensor
Networks with Energy Constraints”, to appear at ICC'05,
[1] S.S., Kulkarni, “TDMA services for Sensor Korea, May, 2005.
Networks”, Proceedingsof 24th International [14] J. Ding, K. Sivalingam, R. Kashyapa, L. J. Chuan, “A
Conference on Distributed Computing Systems multi-layered architecture and protocols for large-scale
Workshops, Pages: 604 – 609, 23-24 March 2004. wireless sensor networks”, IEEE 58th Vehicular
[2] W. Ye, J. Heidemann, D. Estrin, “Medium Access Technology Conference, 2003, VTC 2003-Fall 2003,
Control with Coordinated Adaptive Sleeping for Volume: 3, Pages:1443 - 1447, 6-9 Oct. 2003.
Wireless Sensor Networks”, IEEE/ACM Transactions [15] M. Zorzi, “A new contention-based MAC protocol for
on Networking, Volume: 12, Issue: 3, Pages: 493 - geographic forwarding in ad hoc and sensor networks”,
506, June 2004. IEEE International Conference on Communications,
[3] A. El-Hoiydi, “Spatial TDMA and CSMA with Pages: 3481 - 3485 Vol.6, 20-24 June 2004.
preamble sampling for low power ad hoc wireless [16] R. Rugin, G. Mazzini, “A simple and efficient MAC-
sensor networks”, Proceedings of ISCC 2002, Seventh routing integrated algorithm for sensor network”, IEEE
International Symposium on Computers and International Conference on Communications, Volume: 6,
Communications, Pages:685 - 692, 1-4 July 2002. Pages: 3499 - 3503, 20-24 June 2

6
© 2010 ACEEE
DOI: 01.ijns.01.02.01

Spread Spectrum Based Energy Efficient Wireless Sensor Networks

More Related Content

What's hot (17)

Viewers also liked (9)

Similar to Spread Spectrum Based Energy Efficient Wireless Sensor Networks (20)

More from IDES Editor (20)

Recently uploaded (20)

Spread Spectrum Based Energy Efficient Wireless Sensor Networks