Simulation to track 3 d location in gsm through ns2 and real life

! " #!#
10.5121/jgraphhoc.2010.2103 18
Anand Gupta, Harsh Bedi , MS Don Bosco, Vinay Shashidhar
Department of Computer Engineering
Netaji Subhas Institute of Technology,
Delhi University, New Delhi, India
anand@coe.nsit.ac.in, harshbedi@gmail.com,msdonbosco@gmail.com,
vinayshashidhar@gmail.com
ABSTRACT
In recent times the cost of mobile communication has dropped significantly leading to a dramatic
increase in mobile phone usage. The widespread usage has led mobiles to emerge as a strong alternative
for other applications one of which is tracking. This has enabled law-enforcing agencies to detect over-
speeding vehicles and organizations to keep track its employees. The 3 major ways of tracking being
employed presently are (a) via GPS [1] (b) signal attenuation property of a packet [3] and (c) using GSM
Network [2]. The initial cost of GPS is very high resulting in low usage whereas (b) needs a very high
precision measuring device. The paper presents a GSM-based tracking technique which eliminates the
above mentioned overheads, implements it in NS2 and shows the limitations of the real life simulation. An
accuracy of 97% was achieved during NS2 simulation which is comparable to the above mentioned
alternate methods of tracking.
KEYWORDS
Global System for Mobile Communication (GSM), Network, Global Positioning System (GPS), Signal
Attenuation, Triangularization, Network Simulator, Signal Mast.
1. INTRODUCTION
In modern world, location estimation has almost become a necessity. It helps law enforcing
agencies to keep track of over speeding vehicles [6], and assists organizations in keeping a
check on inefficient employees [8].
Location estimation has almost become a necessity in modern times. Law enforcing agencies
keeping track of over speeding vehicles [6] and organizations keeping a check on inefficient
employees [8] are just some examples where this method has proved to be more than just useful.
GPS, the classical way of tracking, involves determining the latitude and longitude position of
an object with a very high accuracy. However this requires a dedicated device for
communicating with the GPS access points. Thus it increases the cost to make avail of this
service.
Hence in the recent years the focus has now shifted on finding other ways of tracking. This has
led to the emergence of alternate technologies. GSM Network technique is one of them. Use of
GSM network for tracking, in the initial days, did not prove useful because of the lack of
widespread use of mobiles. But of late the number of mobile users has increased dramatically
due to the decrease in cost of mobile communication, as a result of which mobile phones have
become a potent tool for tracking.
There are two approaches to make use of mobiles as a tracking device. One of the approaches
makes use of signal attenuation (energy dissipation) [4] property of a packet. This method

! " #!#
19
makes use of the fact that energy, of the packet, dissipated is related to the distance travelled by
it. However this method requires a highly precise and accurate energy measuring device. This
again increases the cost of the tracking system.
Another alternate technique uses the GSM network. It makes use of the signal masts to
determine the position of the mobile. The GSM network technology works by broadcasting a
request packet to all the towers. The towers on receiving the request packet sends a
acknowledge packet back to the mobile. Tracking the mobile requires use of the time-interval
for the round trip of the packet. Greater the precision of the time-interval, greater is the accuracy
of the tracking. As no extra devices are required for this process it eliminates the cost overheads
required by the other two methods
The organization of the remaining paper starts with the system design and the algorithm in
which we have described how tracking is performed using the GSM network technique. The
calculation and the results which come from the NS2 implementation precede the error
calculation. The paper then covers the real life simulation of the GSM network technique and
reports the difficulties faced during the experiment. The future work section covers our
aspirations, and the direction in which we would like to take the research work. We end the
paper with our findings on the topic and the measures which should be taken to make this paper
feasible for mobile tracking applications.
2. MOTIVATION
The existing tracking techniques suffer from high cost overheads whereas the GSM tracking
technique is quite cost-efficient. This has motivated us to pursue the latter tracking technique
which if implemented on a large scale it could solve many real life problems such as
• Thefts of expensive handsets are on the rise nowadays which results in the loss of
important contacts. An algorithm which can track the mobile without any extra cost to
the pocket is the need of the hour.
• This approach would be helpful in firms wherein it is very important for the seniors
who want to keep a strict vigil on their juniors [8].
• An algorithm which could determine the speed of a vehicle without installing expensive
devices would help the law enforcing agencies [7].
The above-mentioned applications have motivated us to contribute to the GSM-based mobile
tracking which is as follows.
3. CONTRIBUTION
Our contribution towards the GSM tracking algorithm are as follows
a. This paper is an extension to our previous paper titled ‘Accuracy of 3D location
computation in GSM through NS2’ in which we had only simulated the algorithm on
NS2. No real-life analysis was performed.
b. In this paper we are proposing an algorithm that drastically reduces the cost involved in
tracking and locating the position of a mobile user. Even though privacy issue could be
raised the algorithm is self sufficient to tackle the issue

! " #!#
20
c. The algorithm proposed will be a very important tool in the hands of the law enforcing
agencies as they can now easily get to know the whereabouts of a speeding vehicle and
the exact speed at which it is travelling. If could discreetly this would also help
companies keep track of their employees.
d. This algorithm drastically reduces the cost involved in tracking and locating the position
of a mobile user. Even though privacy issue could be raised the algorithm is self
sufficient to tackle the issue.
To continue with the tracking technique we proceeded to simulation of the network in NS2.
4. NS2 SIMULATION
NS2 is an open-source discrete event network simulator. It supports an array of popular network
protocols, offering simulation results for wired and wireless networks alike. It can be also used
as limited-functionality network emulator. Hence, NS2 provides a good platform for simulating
the GSM network and testing our proposed algorithm.
Our algorithm specifies the following protocol.
4.1. Protocols for Location of GSM Mobile User
The workflow diagram for the protocol employed for the simulated algorithm is depicted in
figure 1. The Steps followed are mentioned below.
Figure 1 Protocol used in GSM technique for tracking
1. Tower sends a broadcast signal(with a timestamp field that contains the exact time of
transmission of packet) directions so that nearby towers may accept this signal
2. Towers accept this packet and sends an acknowledge packet which contains the co-
ordinates of the tower as well as the original timestamp and uni-cast it back to the mobile
User request to
start tracing Acknowledge Packet
Acknowledge Packet
1
2
3

! " #!#
21
3. The mobile receives the acknowledge packet and calculates the time interval for the round
trip.
4.2. Assumption
The protocol mentioned in section 4.1 was implemented with the following assumptions
a) The time taken for sending and receiving the signal from the mobile to the tower is assumed
to be so small that user remains stationary in that time frame.
b) The mobile is at any given point of time able to successfully communicate with all the
nearby towers.
c) There is no interference with the signals and the signal from the nearest one reaches the first
and so on.
d) Signal strength of packets is uniform in all directions into which it is directed.
e) During the journey there is no such place which does not come under the region of towers.
f) When 2 or more packets are reaching a tower or a mobile, at any time instant, the mobile or
tower can write the timestamps on the packets together at the same time
4.3. Algorithm
With the above stated assumptions we started the implementation of the algorithm. After
receiving a trigger from the user, Algorithm1 Track_init() initializes the tracking process. It
broadcasts request packets, containing the time of transmission to all the towers in its range .On
receiving the request packet, Algorithm 2 Tower_response() method acknowledges by adding
its co-ordinates to the packet header along with the time of receiving of the request packet. This
uni-cast communication is done between the tower and the mobile from where the request
packet originated. Finally, Algorithm 3 Mobile_coord() method calculates the position of
mobile by using the response packets from first three towers.
Algorithm 1: Track_init()
Input : Tracking request signal.
Output : Broadcasting request packets to the towers.
Method : Mobile broadcast the request packets.
1. while (track_mobile() == NO)
2. ; // wait
3. for I = 1 to N
4. request[i] = new packet()
5. request[i].timestamp = current_time()
6. request[i].mob_id = mobile_id
7. send_tower_request_packet(request[i])
8. end
Algorithm2 : Tower_response ( )
Input : Request packet from mobile.

! " #!#
22
Output : Acknowledge packet from tower containing its coordinate.
Method :Tower responds with a reply packet adding its coordinate.
1. while (received_packet() == NO)
2. ; // wait
3. mobile_id = get_mobile_id_read_packet()
4. old_time_stamp = get_time_stamp_read_packet()
5. acknowledge = new packet()
6. acknowledge.destination = mobile_id
7. acknowledge.tower = tower_id
8. acknowledge.tower_coord = tower_co-ordinates
9. acknowledge.timestamp = old_time_stamp
10. send_mobile_acknowledge_packet(acknowledge)
11. end
Algorithm 3: Mobile_coord()
Input : Acknowledge packet from towers.
Output: Coordinates of the Mobile.
Method: Evaluates the coordinates of the mobile.
1. for I=1 to 3
2. while (receive_ack_packet() == NO)
3. ; // wait
4. time_prop = current_time() – acknowledge.timestamp
5. distance[i] = ((time_prop-delays)*speed_of_light)/2
6. coord[i] = acknowledge.tower_coord
7. calculate_cord()
The following experimental results were obtained on simulating the GSM algorithm.
4.4. Experimental results
The GSM tracking technique is implemented in network simulator (NS2) in accordance with the
protocols mentioned above. The results were collected from the simulation and the calculations
have been shown in the subsequent sections.
4.4.1. Topography used
Figure 2 Depicts a typical mobile cell

! " #!#
23
To enable the GSM based handset the cellular service providers generally divide the region into
small regions called cells, which are generally hexagonal in shape. A typical cellular cell has
been shown in figure 2.
The tower numbering starts from 0 and goes till 5 with the leftmost node being 0 and the rest
numbered in clockwise fashion as shown in figure 2.
TOWER 0: (0, 5, 0) TOWER 1: (0, 15, 0)
TOWER 2: (8.66, 0, 0) TOWER 3: (8.66, 20, 0)
TOWER 4: (17.32, 5,0) TOWER 5: (17.32, 15, 0)
Mobile Position: (4, 3.5, 0)
Figure 3 NS2 simulation results
4.7. Formula Used
Time interval between transmission of packet = a constant delay + time required for and it’s
receiving by tower propagation
Time required for propagation = distance travelled / speed of light
The table given below has been tabulated using the results from the NS2 simulation (figure 3)

! " #!#
24
Tower Time interval Distance
calculated
Actual
distance
%error
0 7.58014*10-4 4.2 4.272 2.20
2 7.58019*10-4 5.7 5.828 1.02
1 7.58041*10-4 12.3 12.176 -1.68
4 7.58044*10-4 13.2 13.404 1.52
3 7.58056*10-4 16.8 17.145 2.01
5 7.58058*10-4 17.4 17.596 1.11
Table 1 Tabulated Result
The calculations which have been performed are shown below
4.8. Calculation result
For Calculation purposes the following equations were used
T1(x1, y1, z1), T2(x2, y2, z2), T3(x3, y3, z3) are the coordinates of the closest three towers
M(x,y,z) are the coordinates of the mobile phone
(x-x1)2
+ (y-y1)2
+ (z-z1)2
= r12
............ (1)
(x-x2)2
+ (y-y2)2
+ (z-z2)2
= r22
............ (2)
(x-x3)2
+ (y-y3)2
+ (z-z3)2
= r32
............ (3)
Subtracting Eq: (1) –Eq: (2) we get
1x + 1y + 1z = 1 ............ (4)
2x + 2y + 2z = 2 ............ (5)
3x + 3y + 3z = 3 ............. (6)
Where i, i, i, i are scalars obtained on subtraction of the equations
Hence,

! " #!#
25
Smallest 3 distance/smallest 3 time delays:
1. 4.2
2. 5.7
3. 12.3
Value found is
X co-ordinate=3.84
Y co-ordinate =3.31
Z co-ordinate =0
Error calculation:
In x- co-ordinate: 3.79%
In y co-ordinate: 5.21%
In z co-ordinate: 0%
5. REAL LIFE SIMULATION
Having completed the NS2 simulation, we proceeded to model the real-life scenario using 4
laptops. The workflow diagram of the experiment has been shown in Figure 4.
Figure 4 Workflow diagram of the real-life modelling of GSM network using 4 laptops
The steps followed are:
1. Simulation was carried out by setting up the wireless ad-hoc network between the 4
laptops treating 3 as towers and 4th
one as the mobile.

! " #!#
26
2. A GSM like network was simulated by exchanging the packets between the laptops
using ICMP protocol.
3. The packets that were exchanged between the laptops acting as towers and mobile were
sniffed from the OSI physical layer using ethereal packet sniffer.
4. The timestamp obtained from the sniffed packets were used to calculate the total
network delay involved in sending and receiving of ICMP packet between the towers
and the mobile.
5. The delay due to the kernel (i.e. the delays involved in the kernel to move from the
application OSI layer to the physical OSI layer) can be determined by self-pinging(ping
127.0.0.1) and determine the time-delays involved.
6. The kernel delays are subtracted from the total network delay to get the propagation
delay which is the time interval involved in the propagation of the packet with the speed
of light.
The observations hence obtained were compared with those from NS2.
5.1. Observations obtained from Ethereal:
The results on pinging from mobile to tower1 have been recorded in figure 5
Figure 5. Ping from mobile to Tower1

! " #!#
27
5.2. Formula used
The hypothesis of GSM tracking proposed by us is based upon the following equations which
are as follow
T = + time_prop .... (7)
Where
be constant internal delays,
time_prop be time involved due to propagation of packet only,
T be the turn-around time between sending of ICMP packet by the mobile
and receiving of the reply packet by the same mobile laptop
distance[i] be the distance travelled by ith packet,
c be the speed of light.
Also, time_prop = D/c.... (8)
Where
D be the total distance of propagation by the packet wave
And c be the speed of light
5.3. Calculations
The calculation were carried out in accordance to the formulae given in section 6.2 and the
observations given in figure 5,figure 6 and figure 7.
S.No
Time Interval between sending and
receiving of ICMP packet (in
milliseconds)
1. 0.783
2. 0.799
3. 0.690
4. 0.985
5. 0.567
6. 0.533
7. 0.671
Table 2. Ping between mobile and tower1

! " #!#
28
S.No
milliseconds)
1. 0.543
2. 0.664
3. 0.764
4. 0.667
5. 3.608
6. 0.674
7. 0.645
Table 3. Ping between mobile and tower 2
S.No
milliseconds)
1. 0.774
2. 0.694
3. 0.714
4. 0.655
5. 0.672
6. 0.778
7. 0.770
Table 4. Ping between mobile and tower 3
5.4. Constraints faced
1. The range of wireless ad-hoc network setup using laptops was
experimentally found out to be 175 metres.
2. The packet captured by the packet sniffer was allotted a timestamp by the kernel of the
operating systems used (Linux kernel version 2.26.x, Windows XP Service Pack and
Windows Vista Home Premium) which is only up to precision 6 digits. So, the delay
which could be calculated only in micro seconds (10-6
seconds).
Maximum Propagation Delay (theoretical)
Maximum range of wireless ad-hoc network
Speed of light
= (175 metres)_ = 0.5833 x 10-6
sec.
(3 x 108
m/sec)
The above calculation shows that the minimum precision required for the timestamp, to
determine the distance between the mobile and the tower using the algorithm specified should
be at least in the order of 10-7
sec.
This time difference cannot be observed by the laptops due to the above mentioned constraints.
6. CONCLUSION
This paper implements network based tracking algorithm both in NS2 as well as real-life
simulation for tracking GSM mobile devices. To the best of our knowledge no one has
simulated this algorithm in NS2. This paper advocates the use of triangularization approach [9]
for tracking GSM mobiles. The approach used in simulation comprises of mainly three stages.
In first stage, the first three responses of packets from towers were recorded. In second stage,

! " #!#
29
time difference between the sending and receiving of packet was calculated. Finally in the last
stage, the mobile location was identified.
The simulation results from NS2 confirm author’s intuition (either tracking employees of an
organization or detecting over speeding vehicles) that GSM based tracking technique will be
best suited for commercial purposes over existing methods.
Further the real-life simulation of the algorithm being limited by the precision of the time-stamp
value as provided by the Operating System Kernel of the laptops(which is in the order of micro-
seconds) is unable to provide the propagation delays essential for triangularization. However, on
being provided with a high-precision timestamp device the author believes that the real-life
simulation performed as prescribed in the paper using 4 laptops will behave in accordance with
the NS2 simulation results.
7. FUTURE WORK
This paper attempts to perform the real life simulation of the GSM network techniques. But due
to the constraints mentioned, the simulation could not exactly determine the accuracy of the
techniques in real life. We are presently focusing on working our way around the difficulties
and trying to come up with a complete analysis that will help us in further understanding the
feasibility of the algorithm in real life.
REFERENCES
[1] R.I. Desourdis, A.K. McDonough, S. C. Merrill, R.M. Bauman, D. A. Neumann, J.A. Lucas, D.
Spector, and D. E.Warren, “GPS-based vehicle tracking with meteor-burst telemetry”, Proc. IEEE
Conf. Military Communications (MILCOM 94), IEEE Press, Oct. 1994, pp.690-694, Vol. 3,
doi:10.1109/MILCOM.1994.473883.
[2] Ding-Bing Lin, Rong-Terng Juang, and Hsin-Piao Lin, “Mobile location estimation and tracking
for GSM systems”, Proc. IEEE Symp. Personal, Indoor and Mobile Radio Communications (PIMRC
04), IEEE Press, Sept. 2004, pp. 2835-2839, Vol. 4, INSPEC Accession Number: 8221818.
[3] H.C. Lee, and S. Junping, “Multilayered Model Strategy for Optimal Mobile Location Tracking”,
Proc. IEEE Symp. Personal, Indoor and Mobile Radio Communications (PIMRC 97), IEEE Press,
Sept. 1997, pp. 1009-1013, Vol. 3, doi:10.1109/PIMRC.1997.627038.
[4] C. Chao-Lin, and F. Kai-Ten, “Hybrid Location Estimation and Tracking System for Mobile
Devices”,Proc. IEEE Conf. Vehicular Technology Conference (VTC 05), IEEE Press, 2005, pp.
2648-2652, Vol. 4, doi: 10.1109/VETECS.2005.1543815.
[5] Security firm using GPS Sacramento Business Journal - June 24, 2005 by Mark Larson Staff Writer
[6] T. Stamoulakatos, A. Yannopoulos, T. Varvarigou, and E. Sykas, “Hidden Markov Filtering with
Microscopic Traffic Modeling for Vehicle Load Estimation in Cellular Networks”, In Proceeding of
IASTED International Conference on Communication Systems and Networks (CSN 2003), Sept. 1-
3, pp. 349-355, 2004.
[7] C. S. Jensen, K. J. Lee, S. Pakalnis, and S. Šaltenis, “Advanced Tracking of Vehicles”, In Proc.
European Congress and Exhibition on ITS, 12 pages (2005).
[8] T. Sohn, A. Varshavsky, A. LaMarca, M. Y. Chen, T. Choudhury, I. Smith, S. Consolvo, J.
Hightower, W. G. Griswold, and E. D. Lara, “Mobility Detection Using Everyday GSM Traces”,
Proc. LNCS Ubiquitous Computing (Ubicomp 06), LNCS Press, Sept 2006, pp. 212-224, doi:
10.1007/11853565
[9] Anand Gupta, Harsh Bedi, M.S. Don Bosco, Vinay Shashidhar, "Accuracy of 3D Location
Computation in GSM Through NS2," netcom, pp.119-122, 2009 First International Conference on
Networks & Communications, 2009.

Simulation to track 3 d location in gsm through ns2 and real life

More Related Content

What's hot (19)

Similar to Simulation to track 3 d location in gsm through ns2 and real life (20)

More from graphhoc (20)

Recently uploaded (20)

Simulation to track 3 d location in gsm through ns2 and real life