COM588 – Internet

Technologies Lecture Notes

By
Kevin Curran
Feel free to use for any purpose.

Wireless Sensor
Networks
 Wireless Sensor Networks
In 1901 Guglielmo Marconi transmitted signals across the
Atlantic from Cornwall to St. John’s, Newfoundland. Throughout
the 20th century wireless transmission has advanced in a
number of different technologies including radio, television,
radar, satellite and mobile.

As sensing, computing and communication are all following an

exponential curve to zero in power, size and cost then it should
make sense that wireless sensor nodes, which are a
combination of sensor, computers, and communicators together
in a small device, are also decreasing in size and cost.

A sensor node is a node in a wireless sensor network that is

capable of performing some processing, gathering sensory
information and communicating with other connected nodes in
the network.

Wireless sensor networks are made up of many nodes, which

are important in their functionality.

The sensor nodes used for wireless networks in particular are

much more adaptable and are capable of coping in more harsh
WSNs also have the ability of wireless communication
over a restricted area. Because of the fact they have
memory and power constraints, they need to be well
organized in order to be able to build a fully functional
network. It is important that the nodes build up the
network using their keys after distribution.

To make sure that they are put in the correct places to

get maximum coverage it is good to plan ahead on where
they are to be placed as this will then allow a plan to be
hatched for the handiest route for them to follow etc and
it is also more secure when the node picks up on more
than one other node as this allows the connections to be
more spread and distribution of information to be spread
further afield in case one failed.

These are ideal for wireless sensor networks as they are

required to be in many types of environments. Sensor
nodes in wireless networks for traffic control are required
to deal with all types of weather.

They are designed to cope with anything from snow to

heavy winds. It is important that they are designed so
they can accurately sense information and can send data.
Evolution of sensor nodes

Figure shows how sensors have decreased in size and

increased in performance over the last 5 years and the
proposed size and frequency of sensors in 2012
(Mainwaring et al., 2002).
Wireless sensor networks are often used for
detecting chemical measurements in industry.
Oil rigs use wireless sensor networks to control
oil amounts and other chemicals. They are
very important and the sensor nodes that are
used have to be adaptable to these
environments to ensure they function properly.

Wireless sensor networks are used to measure

radiation levels. This would mainly be done in
hospitals and power plant industries. Smart
Dust nodes are minuscule micro
electromechanical sensors (MEMS) capable of
detecting anything from light to vibrations.

With breakthroughs in silicon and fabrication

techniques, the size of these motes could
eventually be as small as a grain of sand and
be able to send data to other MEMS up to one
thousand feet.
The military initially used WSN for applications by dropping
sensors in an area from a plane and the sensors themselves
formed the network, which enabled battlefield surveillance to
take place. From here they developed into a cheaper and more
powerful tool.

Sensors are now being placed in very hostile areas where it

would not be practical to install cables. High technology sensors
the size of a grain of dirt are now being used by the military.
These Smart Dust sensors are given a unique ID. The sensors
are aimed at the target from a plane and are inconspicuous as
they also look like dirt. It is believed that sensors contributed to
the death of Al Qaeda leader Abu Musab al Zarqawi.

Sensor Network Applications Development

When referring to wireless devices people normally mention
mobile phones, personal assistants or wireless laptops. Wireless
sensor networks (WSN), unlike these devices, are inexpensive
and do not depend on pre-existing infrastructures.

A WSN is a network made up of multiple independent sensor

nodes, which are usually referred to as motes, equipped with
radio transceivers and a base station which monitor the
surrounding environment for various conditions such as
temperature, speed, motion or gasses.

Formula 1 cars have sensors on different parts of the engine to

monitor key conditions such as the engines performance, the
brake temperature, the oil pressure and the tyre pressure. This
data is fed back to the control centre where the engineers can
see the status of the car on a PC.

A wireless ad-hoc network is normally made up of sensors.

Therefore each sensor operates a multi-hop routing algorithm
meaning that the base station can receive data packets from a
number of different nodes.

One of the advantages of a WSN is that again, unlike the

‘traditional’ wireless devices mentioned above, the more nodes
available the better the network operation. If a large amount of
nodes are in a small location then the network is even stronger
again.
 Network Topologies
There are different network topologies that
apply to wireless sensor networks. The Star
network is a simple communications
topology in which a single base can send
and receive data and messages to a
number of remote nodes.

These remote nodes can only send signals

through the single base and are unable to
send them between each other. The
advantage of this is in its simplicity and the
ability to keep the remote nodes power
consumption to a minimum. The
disadvantage of this is that the base must
be within range with the other single remote
They are unable to operate through each other, therefore
the base is vital. The Mesh network allows any node
within the network to communicate with and through
each other. If a node is out of range from another, then it
can send a message through an intermediate node to
forward the message to the desired node.

The advantage of this network is that if a node fails then

another is available to forward any signal. This network
can also operate in a longer range, where communication
between nodes is generally unlimited.

The disadvantage of this system is that it uses a lot of

power and the more distance between communication of
the nodes, then the longer it will take.

Finally, the Hybrid Star-Mesh network is a more versatile

network and maintains the ability to keep the power of
wireless sensor nodes to a minimum. Nodes with the
lowest power are unable to forward messages, therefore
minimal power is maintained.

However, nodes with more power in this network are able

to forward these messages from those of low power
 Wireless Sensor Hardware
Sensor nodes can vary in size from the size of a glove compartment to a
grain of sand although the latter are yet to be released commercially,
only for military use. The cost of a node varies also, from as little as a
few euros to hundreds of euros.

The price is dependent on the network size and the applications

required. There are generally two types of sensor nodes used in a WSN.
They are sensor nodes and gateway nodes. Sensor nodes gather
required data from their surroundings by means of sensors that are
attached to them. The node links with other nodes in the vicinity to
transmit data to a data gateway.

Sensor nodes need to know about the topology of the network to enable
data to reach the gateway node. Gateway nodes are the interface to the
system or control centre in the WSN. They are also used by the host
when data is being requested from the WSN as shown below.
A sensor node is made up of a power source, a micro-
controller, a transceiver, external memory and sensors
(see below). The sensor is the eyes and ears of the node
and its main purpose is to sense or determine physical
data from the target area.

The sensor generates a signal, which represents a drop in

temperature or the presence of a car for example. The
wave that is generated goes through the analog-to-digital
converter (ADC), which creates a digital signal. The
binary signal is then sent to the micro-controller.

More intelligent sensors such as MEMS have been made

to determine a wider range of activities.
 Transceiver
All sensor nodes have a transmitter and a receiver. These components are both
contained in a transceiver, which enables a node to receive data from another
node and forward the data to another node as necessary. The operational states
are Transmit, Receive, Idle and Sleep. When a radio is operating in idle mode it is
absorbing almost as much power as it would in receive mode. Transceivers do not
have a unique global identifier.

Given the large number of sensor nodes manufactured worldwide it would be

unrealistic to assign one to each node. The radio band used by sensor nodes is the
industrial, scientific and medical (ISM) band that uses un-licensed frequencies.
Components that use the ISM band need to be able to handle interference from
other sources that may overlap. The frequency used by WSN’s are between
433MHz and 2.4GHz.

Communication between medical devices may occupy multiple bands and use
different protocols. To prevent interference in the excessively used unlicensed ISM
band, biomedical devices may use the licensed wireless medical telemetry services
(WMTS) band at 608 MHz.

There are three modes of communication that could be used with sensor nodes.
They are laser, infrared and radiofrequency (RF). As energy consumption is one of
the major flaws with sensor nodes one would expect laser to be the preferred
option as it uses less power than RF, does not need an antenna and is also robust
when it comes to security.

However, laser is very susceptible to atmospheric conditions. It also requires line of

sight, which cannot always be guaranteed given the sometimes-covert nature of
sensor node deployment.

Although infrared doesn’t need an antenna it has a restricted broadcasting

capability. RF, which does need an antenna, is the easiest to use.
 Microcontroller
The microcontroller processes the data that it receives from other
components of the sensor node. Other options include microprocessors,
and field-programmable gate arrays (FPGA). FPGAs are not very power
efficient and do not work well with traditional programming techniques.
They can be reprogrammed and reconfigured as necessary to eliminate
deployment costs but this takes time and energy. Microcontrollers are a
suitable choice for sensor node. They are the best choice for embedded
systems.

Because of their ability to connect to other devices, programmable,

power consumption is less. This is because these devices can go to
sleep state while part of controller is still active. The platform on the
microcontroller needs to contain an instruction effective CPU, different
HW interfaces, memory and a wake-up time.

A power effective processor guarantees fast processing time and low

power consumption. In order to keep expenses low and to prevent any
further components being added to the already limited space the
maximum number of feasible interfaces needs to be on a
microcontroller.

Microcontrollers and FLASH memory are the most applicable memory

types given their amount of memory storage and price. The amount of
memory needed depend on whether the memory is to be used for
storing application data or for programming the device. The memory
should generally be as little as possible as long as there is room for the
application software and protocol. This will ensure that the cost and
 Power Unit
• Power consumption is one of the prime issues WSN have to
deal with. There is more power required for communication than
there is required for sensing or computing.

• The energy cost of transmitting 1 Kb a distance of 100 m is

approximately the same as that for the executing 3 million
instructions by 100 million instructions per second.

• Depending on the surroundings, two types of batteries are

used, chargeable and non-chargeable as it would be impossible
to change the battery in a hostile environment.

• Nowadays, sensors are being developed to recharge using

solar power, thermogenerator or vibration energy. Another
technology being used in sensor power units is Dynamic Power
Management (DPM). This enables devices that are not being
used to be switched off and on as necessary. This is performed
with help from the operating system.

• Another power saving method is Dynamic Voltage Scheduling

(DVS). In this method it is possible to reduce the power by a
factor of four by varying the voltage with the frequency.
 WSN Operating Systems
The operating system of a node is not as
complicated as those of a standard PC. WSN
applications have certain tasks that would not be
used on a standard PC.

User interfaces would be one example that would be

used generally on a normal operating system but
not on sensor nodes.

The main characteristics of the operating system

should be as follows:

•Small in size and energy efficient as sensor nodes are

constrained in memory and power
•Ability to deal with packets in real-time and on-the-fly
•Provide a high degree of software modularity for
application specific sensors
•Should be robust and reliable and able to workaround
individual device failures
 WSN Operating Systems….
As hardware for wireless network systems are the
same as traditional systems, operating systems
such as eCOS and uC/OS can be used for wireless
sensor networks.

TinyOS is the operating system that was initially

designed for sensor networks and was designed to
be used in conjunction with the smart dust project.

What is now known as the TinyOS Alliance started

out as a project between University of Berkeley,
California and Intel Research. It has been
implemented by developers worldwide, on many
platforms, for a wide variety of wireless sensor
networks.

It is written using the nesC programming language,

which is based on the C programming language but
aimed primarily at sensor networks where demands
 WSN Operating
Systems….
TinyOS is free and is used on the MICA2 platform.
The fact that it is open source also means that many
of the software components needed to use the
platform are already written. This also makes it
easier for the software to interface with different
hardware.

It is an event-driven operating system in that it is

driven by external factors detected by the sensor.
The core of the operating system is about 400 bytes
which makes it suitable for every type of modern
micro-controller. TinyOS has a two level scheduling
structure i.e. its ability to perform extensive tasks
which can be interrupted by an event.

There are two interfaces, command and event. To

smooth the progress of modularity, each component
declares the commands it uses and the events it
 Conclusion
•A sensor node, also known as a mote, is a node in a wireless sensor
network that is capable of performing some processing, gathering
sensory information and communicating with other connected nodes in
the network.

•Wireless Sensor Networks are ideal for monitoring and analysing. Many
shops use wireless sensor networks as a means to stop thieves. These
sensors are equipped with alarms to ensure security levels are kept to a
maximum.

•Wireless sensor networks are also suitable for monitoring movements

in urban areas such as the desert and forests. It allows these areas to be
monitored without the expense of labour.

•Wireless sensor networks are used to try and detect the risk of natural
disasters as soon as possible and are fast to send data, giving people a
chance to act. Security, using wireless sensor networks is one of the
most important areas where it must be working correctly.

•In more urban areas there is less time dedicated to maintenance.

Power must also be available to all security sensors at all times. A failure
in power can result in high consequences if timing goes wrong.

•Some areas to indicate where wireless sensor networks can be used for
security issues include monitoring disease, monitoring floods, vehicle
tracking, monitoring volcanic eruptions, alarm signalling and forest fires.