Wireless Sensor Network for Remote Monitoring

and Detection of Explosives

(W-ReMADE)

Simi S. and Maneesha V. Ramesh

Amrita Centre for Wireless Network and Applications, Amrita School of Engineering

Amrita University, Ettimadai, Coimbatore, India

Abstract. Recent years have shown a worldwide increase in terrorist bombings.

Continuous monitoring for the presence of explosives in public places can

improve security of the public and infrastructure. The objective of this research

work is to reduce, control, and warn about the forthcoming terrorist activity by

precise and quick detection of explosives. This paper proposes a wide area

monitoring system using a multi phase wireless sensor network design. W-

ReMADE uses multiple wireless sensor nodes integrated with different types of

sensors to identify the explosives. Based on diverse orthogonal techniques, the

system collects data from the sensing nodes, dynamically aggregates the data

and forward to the sink node for further analysis. A mobile node has been

introduced to further confirm the suspected objects, thus offering an enhanced

target tracking mechanism that reduces number of false alarms. W-ReMADE

provides an effective warning mechanism for security threats in public places so

that immediate action can be taken against bomb threats.

1 Introduction

In today’s world terrorism is a main threat to the security of the world. Continuous

monitoring systems can improve the security of infrastructure and general public in

urban areas. A wireless sensor network is a solution for continuously monitoring and

identifying explosive materials. Currently, effective systems were not developed to

operate remotely in open environments over a wide area. The difficulty with the

existing techniques is that the suspected items have to bring nearer to the detecting

instrument. This involves more human involvement in the detection and cannot be

continuously monitored. Here comes the significance of remotely detecting the

explosives were the process of detection is taking place at a reasonable distance from

the suspected material without affecting the other people occupied in suspected area.

This paper mainly concentrates on detection of explosive materials in public

places with the help of wireless sensor network technology [1-2]. The area under

study is monitored in real time, collect data, aggregate it and send to the sink node.

The main constituent of explosives is chemical components. By analyzing the

chemical signatures, it is almost possible to predict whether a material is explosive or

not. In an open environment, a single type of sensor may not be adequate in

confirming the explosive material presence. Also the concentration of these materials

S. S. and Ramesh M..

Wireless Sensor Network for Remote Detection of Explosives.

DOI: 10.5220/0003116300600071

In Proceedings of the International Workshop on Semantic Sensor Web (SSW-2010), pages 60-71

ISBN: 978-989-8425-33-1

 2010 SCITEPRESS (Science and Technology Publications, Lda.)

will be very less in the atmosphere because of its well packing. These may cause

wrong alarms and destroy the usefulness of the system. To meet this limitation, the

W-ReMADE uses more than one mutually independent technology for the detection

scheme. As the signal strength is less in field, the system uses a mobile node to reach

the suspected area, collect signals and confirm the presence of explosives and thus

provide an enhanced target tracking. Sensor nodes of the wireless sensor network are

planned to deploy in Indian airports. Multiple sensors of varying type are deployed in

such a way that the network can cover the whole region. Sensed data is send to the

sink node for processing.

According to global terrorism data base [3], the terrorist attacks are increasing in

recent years. In terms of total terrorist attacks between 1970 and 2007, India is ranked

in the fifth position in the list of top ranking countries. It is also found that nearly 50%

of weapons used were explosives. The explosives used were readily available,

especially dynamite, grenades, and improvised devices placed inside vehicles. The

security measures in airport are not much effective to detect the presence of

explosives. W-ReMADE is designed to set up not only in the inside areas of airports,

but also to provide an effective way to identify the presence of explosives outside the

rooms. Even if it is designed for this particular scenario, this can be applicable to

other public places such as railway stations, bus stations, parks, embassies, hotels etc.

with slight design modifications.

The paper is organised as follows: Section 2 describes a brief review of the related

works. Section 3 and 4 presents the proposed system model and network architecture.

The algorithms used in the system are described in section 5. Section 6 deals with the

advantages of the system followed by the conclusion.

2 Related Work

In conventional systems, biosensors like dogs and honey bees were used to detect

explosives. But they have restricted attention span and are very expensive. So, various

instruments have been developed. The problems with those approaches were the

complexity and bulk size. Several existing detection methods that can be utilized for

remote explosive detection are mentioned in [4]. The development of explosive

detection with Micro Electro Mechanical Systems(MEMS) technology was briefly

reviewed in [5] MEMS is a micro fabrication technology that combines mechanical

elements, sensors, and electronics in a chip. For commercial application of potential

MEMS based explosive detectors, require high sensitivity and excellent selectivity.

Here also standoff distance is a main problem to apply in open environments.

In [6] authors developed a technique for stand-off detection of trace explosives

using infrared photo-thermal imaging. They used a set of infrared quantum cascade

lasers which is tuned to the absorption bands of explosive traces .When the lasers

illuminate the object, an infrared camera detects the small increase in thermal signal.

The main problem with this approach is the standoff distance limit. Also in a noisy

environment, it is very difficult to detect the increase in thermal energy. The authors

of [7] utilize terahertz technology for explosive detection. The system uses very low

levels of non ionizing radiation to detect and identify objects hidden under clothing.

Many chemical substances and explosive materials exhibit characteristic spectral

responses at THz frequencies that can be used for threat object identification. This

technique is able to sense through several layers of clothing with the help of safe non

ionizing radiations. As the maximum standoff distance that can be achieved from this

method is 1m, in an open environment it is difficult to apply this method.

Some of the stand-off methods currently developed is focused on chemical

identification. The main challenge includes the distances from which effective

detection can be conducted in presence of various interferences from environments.

Bourzac, Katherine describes a method to detect explosive materials using magnetic

sensors developed for use in the battlefield [8]. The National Institute of Standards

and Technology (NIST) have developed magnetometer for detecting the presence of

magnetic materials [9]. But it does not consider the information about the chemicals

used in explosives and it cannot be applied to scenarios where more metallic presence

is found. German researchers developed a sensor system [10] to monitor people

carrying explosive in public places. The system consists of two separate sensor

networks to find chemical properties and kinetic information of the person. They are

using their own chemical sensors for the sensor network. The cost of developing such

type of sensors is comparatively high. In the proposed design, the system uses already

existing components for sensing purpose. As one type of sensor is not sufficient to

detect the explosive presence, the proposed design utilizes more than one independent

technology. Also W-ReMADE allows to continuously monitoring the area without

affecting the passengers going through the monitoring area.

3 W-ReMADE System Design

3.1 Sensing Components

As the system is using more than one technique is used to identify the explosive and

the joint result is used for decision making, the probability of false alarms are

comparatively less. Even if one of the sensor readings is wrong, the system can work

well by selecting the correct reading from other sensors. An explosive material can

be identified chemically, magnetically, thermally and electrically. But thermal and

electrical measurements will not improve performance because of the noisy

environment. So the proposed design utilizes the chemical and magnetic properties

of the material. W-ReMADE combines imaging technology, optical technology and

chemical as well as magnetic identification techniques.

In W-ReMADE, we used a set of image sensors to locate unattended objects.

Theses image sensors will capture the pictures of the scene periodically and send to

the image analyzing server to identify unattended objects. To support chemical

identification of the explosives, W-ReMADE makes use of vapor sensors to collect

the vapour concentration present in the air. An air collecting system and a filter is

used to collect large volumes of air and to filter the commonly present particles of

air. The vapour sensor is connected to the mote using an interface board and it can

transmit the sensed data for analysis. To magnetically identify the explosives, the

system used the MicaZ magnetometer. The system also employs an optical

technique in which it uses a LIDAR emitting infra red radiations that will excite the

molecules of air and reflected back. To collect the response patterns, the system is

equipped with additional light collectors and detectors. The corresponding

frequencies of vibration are calculated and send to the sink node for further analysis.

3.2 System Architecture

The proposed wide area monitoring system is applicable to any public places such as

railway stations, airports, bus stations, supermarkets, embassies etc. Even if the

system is applicable to all urban areas, this project mainly focuses on the airports in

India. The area under testing is equipped with LIDAR, image sensors, magnetic

sensors and chemical sensors. These sensing components are deployed in the roof of

the passenger areas. The system will continuously monitor the area and if the strength

of signal collected from the sensors is greater than a particular threshold, the system

will immediately give indication to the security personnel through internet or mobile

network. If the collected signal strength is less than the predefined threshold, the

system will perform second phase operations for the confirmation. The following

figure 1 gives an overview of the system.

Fig. 1. Overview of the system.

Phase 1

To monitor the presence of explosives, W-ReMADE uses vision based method,

chemical identification and magnetic identification. All these methods are done in

parallel in order to provide more precise results and to reduce number of false alarms.

The system will compares the data collected from different techniques and make an

appropriate decision whether any of the data is false or not. Even if the result from

one technique is wrong, the system is able to distinguish it by using the correlation

between historical data and data from sensors using a different technique. Image

sensors will periodically take pictures of the scene and send to the image analysing

server. The background images of the test area are already stored in the server. By

running the object identification algorithm, the system is able to find out unattended

objects. If any unattended object is found, the area under that image sensor will be

close monitored by initiating phase II operations for the confirmation of suspected

object.

The concentration of explosive molecules may be low in the test environment. To

get more concentrated vapors of the air, W-ReMADE uses a vacuum system which

will collect large volume of air from the test area. These air molecules are filtered

using a concentrator and fed to a vapour sensor. From the concentrated air molecules,

the vapour sensor can effectively find the chemical composition of particles present in

that air sample. The sensed data is send to the sink for analysis. Using a

magnetometer, the system will find the metallic presence in the area. The laser beam

from the LIDAR scans the entire area with the IR beam and excites the particles

present in air. The system is equipped with additional light collectors and detectors to

capture the response patterns. These response patterns are fed to a vibration sensor

tuned to the resonant frequency of chemicals in the explosive. From the vibrations of

the sensor, the system will find the corresponding wavelength. These detected

wavelengths are forwarded to the sink for further analysis. Sink node contains a data

base that stores the chemical signatures of the already known chemicals present in the

explosives. Sink node will compare the received chemical signature with one in the

data base and take appropriate decisions by considering all other measurements.

The system uses two thresholds to process strong and weak signals. A strong

signal is identified by a high threshold t

high

and to discard the weak signal the system

uses a low threshold t

low.

If the sensed data is greater than the high threshold t

high

, it

will give immediate indications to the security personnel’s. If the comparison result is

less than t

high

and greater than t

low

, it will initiate phase II operations for the

confirmation of presence or absence. If the calculated result is less than t

low

, the

system will ignore the data. .Figure.2 gives the design details of W-ReMADE.

Fig. 2. W-ReMADE system design.

Phase II

Operations in phase II will be initiated whenever the sink node receives some

suspicious data. Then sink node initiates a mobile node with more sensitivity for close

observation. A special track made of steel rope is provided in the roof for the

uninterrupted movement of mobile node. To deploy such a path is cheap in test area

like airports. This node will reach the suspected area and scan the area for explosives.

The sensed data is immediately sent to the sink for verification of suspicious data. If

the sink node confirms the presence of explosives, it will give indications to the

security personnel through e-mail or internet or sms services so that they can take

immediate actions. This mobile node can also be utilized, if any anonymous phone

call about bomb threats is received.

4 W-ReMADE Algorithm Design

4.1 Network Topology

For efficient routing of packets within the network, the entire area is divided into a

number of clusters. Based on the communication range, the sensing nodes will

create clusters. The proposed design uses a hierarchical architecture. In a cluster,

there are multiple sensors of same type. Even if the data collected from one sensor is

erroneous, the system can effectively calculate correct results by using data from

alternate sensor. W-ReMADE will correlate the data coming from the sensors in a

particular area, find the deviations of data and can ignore data accordingly. This

allows the system to reduce wrong data processing. Also if any one of the sensor is

not working properly, the remaining set of same type sensors can contribute data and

can manage sensor faults. The following figure 3 illustrates the network topology.

Fig. 3. Network Topology.

There are mainly four types of nodes in a cluster. These nodes are cluster member

node, cluster head, temporary cluster head and connection node. The low level nodes

which sense the data are called cluster members. Each cluster is associated with a

special node with more computational capacity called cluster head. The main

functions of these nodes are to aggregate the data coming from cluster members and

to forward to sink node. A small number of alternative cluster heads are provided to

the network which can be shared among clusters to handle cluster head failures if any.

Connection nodes in the network act as communication link between two clusters.

4.2 Algorithms

Cluster Generation Phase

This phase mainly focuses to generate clusters of sensor nodes for effective

communication. The algorithm will select one of the available special nodes as cluster

head for each region. Initially, in each cluster head, the administrator of the network

will load the number of hops from that cluster head to the sink node. This is for

computing the shortest route to the sink node from each cluster head. To set up a

cluster, the nodes in the network will exchange a set of messages. These messages are

invitation message, response message, confirmation message, negotiation message

and acknowledgement. Invitation message is a broadcast message used by the cluster

heads to invite other nodes in its communication range to create clusters. This

message contains the ID of the cluster head. Response message is send by the nodes

who receives invitation message. This is an indication that the node is reachable from

the cluster head and it is ready to join the cluster. The message includes the node ID,

number of invitations, and the IDs of inviting nodes and corresponding signal

strengths. Negotiation message is transferred between the cluster heads to

compromise the number of cluster members and link nodes in each cluster. This

maintains a minimum and maximum limit in the number of nodes in the cluster.

Confirmation message is send by the cluster heads to confirm the membership in the

cluster by specifying the ID of the cluster head. After receiving confirmation

message, the cluster members will send an acknowledgement to the cluster head.

In the cluster generation process, cluster heads will broadcast an invitation

message to all the neighboring nodes. The nodes receiving invitation will send a

response message to all the inviting nodes. If the signal strength of any of the

invitation message is less then it will ignore the invitation otherwise send a response.

The cluster head will store the details of response messages in a table. The nodes

receiving invitation from more than one cluster head are the candidate of a connection

node which is a bridge between the communications of two cluster heads. Using any

one of the connection nodes, the neighboring cluster heads will communicate with

each other to make an agreement between numbers of cluster members and

connection nodes. The network design supports only at most two connection nodes

between two clusters in order to avoid energy wastage of these nodes. Depending on

the total number of cluster members of neighboring cluster heads, the extra

connection nodes will be changed to cluster members of any one of the cluster head

and update the cluster table. The cluster heads will send a confirmation message to all

nodes in its cluster table. By receiving this confirmation message, the member nodes

will store the ID of cluster head and send an acknowledgement to the cluster head. If

the cluster head is not receiving the acknowledgement after the timeout period, it will

retransmit the confirmation message. All cluster heads in the network knows the

number of hops to the sink node from that node. These cluster heads have to forward

the aggregated data to the sink node. For fast and effective forwarding, the number of

hops travelled by the packet should be less. We used Dijkstra’s shortest path

algorithm [11] to find the shortest path from each cluster head to sink node. It uses

number of hops to the sink as metric of the algorithm. The shortest path information is

added to the routing table of each cluster head and connection node.

Communication Phase

In this phase, the actual communication between nodes takes place. The cluster

members will send the sensed the data periodically to the cluster head. The sampling

rate can be changed by the sink node in case of any suspect in that particular area.

Cluster heads aggregates the data coming from different sensors and forwards it to the

sink. The sink will receive the aggregated message from all cluster heads and analyze

the data. The messages exchanged in this phase are synchronization message, data

message containing sensed data from cluster members and aggregated data message

that is to be forwarded to the sink node.The cluster members will send the data

message to the cluster heads. Cluster heads will check the message type when it

receives a packet. If the message is a data message from one of its cluster member, it

will store the data for further processing, aggregate data, create a packet destined for

sink, find the shortest path node to sink, and forward the data. If the message type is

an aggregated message to the sink, the node will find next shortest path node to the

sink and forward the aggregated message. If it is a synchronization message, the node

will reset its clock and forward to neighboring cluster heads.

Data Aggregation Phase

All the cluster members will collect data periodically and send to the cluster head.

The main function of cluster head is to aggregate the data. There is multiple numbers

of varying types of sensors in a cluster. Based on the timestamp, the cluster members

will store the data from all type of sensors and create a vector. The number of such

vectors in cluster heads will be different based on the number of sensing nodes in a

cluster. It will compute the correlation between each vector using the following Karl

Pearson’s correlation coefficient.



n∑x

 ∑x∑





∑



  ∑





∑



  ∑



(1)

If r>=0.8 or r <=-0.8, then there is a strong correlation between the vectors. So it is

not required to send all the incoming data, only one vector is sufficient. This will

reduce the communication cost of the system. If the correlation coefficient is less than

0.5, it is required send the differing vectors. Cluster head will add a time stamp and

forward the aggregated vector to the sink by selecting the shortest route in the routing

table.

Data Analysis Phase

The chemical and magnetic signatures of already known explosives are stored in the

database of sink node. In the sink node, the received aggregated data is compared

with the signatures stored in the database. If the recieved data is less than the

threshold t

low,

, the system will ignore the data. If the incoming data is in between t

low

and t

high

, the sink node will initiate a mobile node for the confirmation of explosives in

the suspected area. If the data is greater than t

high

, then the system immediately give

indication to the concerned people.

Advanced Sensing Phase using Mobile Node

If the output of data analysis phase is in between t

low

and t

high,

we cannot surely say an

explosive content in the area and cannot disseminate an alarm. It may be an

erroneous/noisy data. In this situation, W-ReMADE uses an enhanced target tracking

phase to confirm the presence or absence of explosive and to avoid wrong alarms. The

sink node will localize the area using the node ID of incoming packets. The area

under monitoring is equipped with a steel rope on the roof. The sink node will initiate

a mobile which can traverse though this path and can reach the area under suspect.

This node is equipped with high sensitivity components which carry out the detection

process more close to the target. The mobile node will send the sensed data

immediately to the sink node for the confirmation of suspected target.

Localization of Suspected Area

In case of suspicious data, the sink node will initiate a mobile node for the confirma-

tion of suspected object. The sink node will look up the distance to the corresponding

cluster head from where the suspected data is obtained. It initiates the mobile node fed

with the location information. Also it sends a message to the cluster head to indicate

that, the sink node initiated a mobile node and it will reach the cluster head with in t

seconds. The cluster head knows the location of static cluster members. It checks the

previous packets and finds the coordinate position of the sensor nodes from which the

suspicious data was obtained. Then it uses triangulation technique to find the

approximate location of suspected target. After t seconds, the cluster head will

broadcast a message with the location information, which can be received by the

mobile node. By receiving this message, the mobile node is able to go more close to

the suspected object. It will sense the data forwards to the sink node and provide a

better tracking mechanism.

Alarm Dissemination Phase

If the presence of explosive material is confirmed by the sink node, then the system

will provide an early warning to the concerned persons. For the indication of

explosive material presence, it uses existing mobile network and internet. The system

will automatically give sms alerts and e-mail alerts to the important security officials.

The authorized persons can view all the sensed data from the sensor network in the

internet in real time. Depends on the variations in the sensed data, the officials can

take immediate actions. In the case of threat messages or calls, the administrator can

configure the system to change the sampling period and threshold limits so that the

system can provide improved results.

Time Synchronization

In a sensor network, there may be propagation delay of the packets due to some

environmental factors. The sensors have to coordinate their actions for the

aggregation of data. If there is no time synchronization, the aggregated data may be

an erroneous. Each sensor node is associated with a clock based on its oscillator

frequency. Due to atmospheric conditions such as temperature, pressure, there may be

slight difference in the oscillating frequency and in turn result in a drift from original

clock. But the network protocol requires a common clock to avoid erroneous data. In

W-ReMADE, The clock in all the nodes of the network is synchronized with respect

to the clock of sink node. Sink node uses a spanning tree algorithm to find connected

components of network graph. It will send a synchronization message with current

clock time to the cluster heads. To handle the difference in clock value due to the

delays in the network, we calculated an estimate of delay of packets. Here we

considered only the transmission delay and propagation delay. Also it is assumed that

the distance between two communicating nodes is a constant r. The propagation delay

between two communicating nodes depends on the distance between them and the

signal propagation speed. The propagation delay (prop_delay) is computed as the

ratio of distance between nodes to the speed of light. Also we calculated the

transmission delay, trans_delay as the ratio of number of bits to the transmission rate.

prop_delay = r / c. (2)

trans_delay = packet length / transmission rate. (3)

delta = prop_delay + trans_delay. (4)

The cluster heads knows the number of hops required to reach the sink node.

Whenever the cluster head receives such a synchronization message, it will multiply

the number of hops and the delay factor to calculate propagation delay. This

propagation delay will be added to the incoming clock data and the local time will be

updated. After the synchronization of cluster heads, the will create a synchronization

packet with updated data and broadcast to cluster members. They will add the delay

factor and update the clock time.

4.2.1 Selection of Frequency of Transmission

As the sensed data in W-ReMADE is more sensitive, it is required to prevent the

unauthorized capturing of data in flowing through the network. To provide security to

the data transmission, we are using frequency hopping [12]. The available bandwidth

is divided into a number of bands. The network operates between 2400MHz and

2483.5 MHz. It is divided into 12 nonoverlapping channels. The channel allocation

algorithm randomly selects a particular channel and used for transmission. The

receiver is also using the same algorithm and seed for the generation of frequency. As

the algorithm randomly selecting the channel, it is very difficult for the intruder to

find the sequence of frequencies used for transmission and thus provides security for

the transmission.

5 Implementation

To implement W-ReMADE, we used MicaZ motes and ZigBee technology. The

sensor nodes and gateway in the system are using CC2420 RF transceiver. Each

cluster member in the system can sense the data and communicate to higher level

nodes. Also they can receive synchronization messages and other control messages

from higher level nodes in the hierarchy. The transmission and reception of the

messages are through MicaZ mote embedded with ZigBee compatible RF transceiver

.It uses a communication frequency between 2400MHz and 2483.5 MHz. With the

help of MicaZ expansion connector, the system can connect to other sensors, data

acquisition boards and gateway. MIB600CA is used to connect the wireless network

to wired network for streaming the sensed data to the internet. TinyOS is the

operating system used for the development. The components and interfaces of

TinyOS are used to communicate messages in the network.

To evaluate the performance of the proposed system, we simulated the

functionality in National Instruments Lab View – Real Time software. For each type

of sensor, we plotted the received data. Due to atmospheric interference, there may be

noise in the collected data and it may cause wrong alarms. To reduce number of

wrong alarms, W-ReMADE uses a confirmation phase where a mobile node can

move closer to the suspected object and can sense data. So the data is more accurate

compared to the stationary sensor data. We assumed that the noise in the stationary

sensor data follows a Gaussian distribution. Due to error factors, the stationary sensor

data crosses above or below the predefined thresholds causing wrong alarms. As the

mobile sensor is getting more accurate data, the probability of wrong alarms are

reduced. The following graph in Figure 4 shows that the probability of wrong alarms

is significantly reduced when we use the confirmation phase. We created a test bed in

a closed air conditioned room. Sensors are placed on the roof and at the air exhausts.

The sensed data is forwarded to the sink node and can be viewed in the internet. We

gradually changed the metallic and chemical concentration in the room. Then the

network was able to give results which exceed the threshold value. A sensor node was

placed more close to the target. The data obtained from that sensor was more exact

compared to the other readings.

Fig. 4. Performance improvement using confirmation phase.

6 Advantages

The W-ReMADE system is applicable to any wide and highly populated areas. The

detection process involves continuous monitoring with minimal human involvement

and without affecting the routines of passengers. As the system uses a confirmation

phase and data from more than one mutually independent technique, the probability of

wrong alarms are significantly reduced. With the help of frequency hopping the

network provide security to the data available in the network. The security officials

need not visit the site for getting details; instead it is available in the network. W-

ReMADE provides facilities for authorized persons to view the sensed data through

the internet in real time.

7 Conclusions and Future Scope

W-ReMADE utilizes wireless sensor network technology to detect explosive

materials present in urban areas. This is a wide area, continuous, remote monitoring

system with minimal human involvement. The system can be deployed to any public

places such as railway stations, airports etc. for the safety of general public and

infrastructure. The accuracy and reliability of the system is maximized, and false

alarms are reduced, by the use of multiple sensors of varying types ensuring coverage

of the wide area. Enhanced target tracking is achieved by using a confirmation phase

with the help of a mobile node. If the system confirms the presence of explosives, the

concerned people are informed via the existing mobile network and internet. One of

the future developments of the system is in reducing the high degree of noise, in

outdoor environments, that can cause performance degradations.

References

1. Akyildiz I, Su W, Sankarasubramaniam Y, Cayirci E, A survey on sensor networks, IEEE

Communications Magazine, 43(5), 102–114, 2002.

2. Pottie G J, Kaiser W J. Embedding the Internet: Wireless integrated network sensors.

Communication of the ACM,,43(5), pp.51~58, 2000 DOI= http://www.ee.ucla.edu/~pottie/

papers/nae_01.pdf.

3. Gary LaFree and Laura Dugan “Introducing the Global Terrorism Database,” Political

Violence and Terrorism 19:181-204. 2007 DOI =http://www.ccjs.umd.edu/faculty/

userfiles/23/FTPV_A_224594.pdf.

4. National Research Council, Existing and Potential Standoff Explosives Detection

Techniques The National Academies Press, Washington, D.C., 2004.

5. Deyi Kong “MEMS Based Sensors for Explosive Detection” in Proc. 3rd IEEE Int. Conf.

On Nano/Micro Engineered and Molecular Systems January 6-9, 2008, Sanya, China.

6. Robert Furstenberg “Stand-off Detection of Trace Explosives by Infrared Photo-thermal

Spectroscopy” DOI =http://

ieeexplore.ieee.org/iel5/5159754/5168000/05168074.pdf

7. C. Baker, T. Lo, W. R. Tribe, B. E. Cole, M. R. Hogbin, and M. C. Kemp Detection of

Concealed Explosives at a Distance Using Terahertz Technology Vol. 95, 0018-

9219/_2007 IEEE No. 8, August 2007 Proceedings of the IEEE 1559-1565.

8. Bourzac, Katherine. “Tiny, Sensitive Magnetic-Field Detectors: Arrays of cheap magnetic

sensors could detect improvised explosive devices.” Technology Review, Massachusetts

Institute of Technology 16 November 2007.

9. Subpicotesla atomic magnetometry with a microfabricated vapour cell DOI =

http://tf.nist.gov/ timefreq /general/ pdf/2219.pdf.

10. “Hidden sensor network to detect explosives” DOI=http://www.theengineer.co.uk/news/

hidden-sensor-network-detects-explosives/1000515.article.

11. Thomas H.Coreman and Charles E. Leiserson , Z ( 2001) Introduction to Algorithms(2

ed.) London: MIT Press and McGraw-Hill.

12. William stallings (1981). Data and computer Communications (2

ed.), mexico: Prentice

Hall.