Real Time Position Location & Tracking (PL&T) using Prediction Filter and

Integrated Zone Finding in OFDM Channel

SHAKHAKARMI NIRAJ
DHADESUGOOR R. VAMAN
Department of Electrical and Computer Engineering
Prairie View A & M University (Member of Texas A & M University System)
Prairie View, Houston, Texas-77446, USA
e-mail: niraj7sk@yahoo.com, drvaman@pvamu.edu, http://cebcom.pvamu.edu

Abstract: - The nature of pre-determined and on-demand mobile network fabrics can be exploited for real time
Position Location and Tracking (PL&T) of radios and sensors (nodes) for Global Positioning System (GPS)
denied or GPS-free systems. This issue is addressed by a novel system of integrated zone finding and
triangulation method for determining the PL&T of nodes when mobile network fabrics are employed based on
using directional antennas for radio communications. Each mobile node is switched dynamically between being
a reference and a target node in PL&T operation to improve the PL&T accuracy of a target node. This paper
presents the Baseline PL&T with predictive Kalman filter and Integrated Zone based PL&T algorithm design
that integrates zone finding and triangulation method. The performance of the proposed algorithm is analysed
using Interleaving-KV sample coding & error correction in Rayleigh and Rician channel using Orthogonal
Frequency Division Multiplexing (OFDM) system under the severe multipath fading.

Key-Words: Real time, Position Location & Tracking (PL&T), Prediction Filter, Integrated Zone Finding, Orthogonal
Frequency Division Multiplexing (OFDM), Channel.

1. Introduction
Mobile ad hoc network architectures can be flexibly residing prior to using triangulation for deriving
deployed and the nodes are highly mobile to the Position, Location and Tracking (PL&T) of the
facilitate supporting a wide variety of emergency target node. Universally, tracking of any device in
disaster scenarios. In some instances these nodes networks has been well established by the use of
can be air dropped and configured into a set of Global Positioning System (GPS). However, the use
clusters and allow immediate network operation to of GPS is not secured and in some instances, the use
support multi-service data exchange. However, of GPS can be denied. Also, GPS cannot work
these network architectures tend to be bandwidth accurately indoors or near to the buildings and it
and resource constrained. They need to be managed cannot detect the devices (also referred to as radios)
skill fully so as to minimize the power of processing in multi-floor buildings. Therefore, the tracking
and overhead transmission to extend the life time of algorithm needs to rely on the use of reference
nodes and allow maximum bandwidth usage for radios to find PL&T of a target radio(s). Three
supporting end user applications. Also, while reference devices are required to track a target radio
preserving the life time of the nodes, it is important in (X, Y) plane using triangulation. Similarly, four
to consider minimization of transmission power to reference devices are required for tracking a target
support a target data rate between peer-to-peer radio in (X, Y, Z). Since the radios move, the
nodes. The use of directional antenna enables the accuracy of tracking suffers due to multiple moving
power to be focused over the particular zone to target locations of the radio. Thus, the triangulation
provide longer range compared to that of using method uses re-initialization of the radio location
Omni-directional antenna [1-3], [6]. We exploit the with a known location by moving the target to that
focused coverage of directional antenna to allow location to improve the accuracy. Also, if the
detection of the zone in which the target is reference radios move in the network, the reference
radios also suffer from accuracy of their locations.
N. Shakhakarmi and D.R.Vaman are with the ARO Center for Thus, in a real network centric operation, radios are
Battlefield Communications (CebCom), ECE Department, Prairie View
A&M University (Texas A & M University system), Prairie View, needed to be switched dynamically to act as
Houston, Texas-77446, USA.
reference radios or target radios at different nodes. Neighbouring radios are used as references to
instances of time. In addition, each disaster scenario track a target radio by using the directional coverage
is different from the other and it is important to of the directional antenna’s beam and the angle of
establish whether fixed reference radios are used or arrival of the directional beam to compute the
moving reference radios are used. Also, the distance (or range) of the target radio from the
feasibility of using only external reference radios for reference unit and the angle of the beam between
tracking indoors or a combination of outdoor-indoor the reference and the target with respect to the
reference radios for tracking indoors. Based upon baseline of a pair of reference radios [3]. Once the
the decision, it is possible to determine whether PL&T location of the target is determined, a
Dynamic Switching of Radios (DSR) between being Kalman filter is employed recursively to predict the
a target radio and being a reference radio for next PL&T locations based on error covariance that
deployment [1-2]. DSR model requires a little bit computes Kalman gain and determine the corrected
more processing power and overhead time for the true position of the radio and the covariance error
management function. [4-7]. These true positions are translated into
This paper describes a novel Position, Location and simultaneous localization and map building based
Tracking (PL&T) algorithm based on Time of on constrained state estimation algorithm [9]. The
Departure (ToD) and Time of Arrival (ToA) recursive prediction is continued until the target
measurements for each Internet Protocol (IP) packet radio goes out of tracking range.
exchanged between a reference radio and a target When GPS is available at each radio, the tracking of
radio for determining the range between a reference the target radio is simpler as it provides GPS data
radio and a target radio. Also using multiple after silent mode. For this case, once the transmitter
references with known PL&T, the range is finds the target, it forms a directional beam and
translated to [X, Y] co-ordinates for 2D tracking and transmits a Request to Send (RTS) message to the
[X, Y, Z] co-ordinates for 3D tracking using target. The target sends a Clear to Send (CTS)
triangulation. It uses KV Transform Coding which message as well as GPS data to the transmitter after
is based on orthogonal transformation of four it completed the data transfer and goes into silent
discrete samples into four coefficient samples for mode. Target receives all of the data from the
transmission. Each discrete sample is creating using transmitter and transmitter performs the Location
n-bits of input data and when transformed, it Prediction Algorithm using Directional
produces each coefficient samples that can be Communication (LPADC) only in the forward
transmitted with 4 bits using any digital modulation direction [6]. The limitation of this approach is the
technique. An ensemble of blocks (referred to as unknown silent mode duration and the ability to stay
KV blocks) with each having four discrete within the coverage area to get the necessary to send
coefficient samples each carrying n-bits are created GPS data. To address this limitation, researchers
for transmission. The ensemble of coefficient have allowed the system not to send CTS until the
samples of M-KV blocks is transmitted to the receiver is in the coverage area and has the ability to
receiving side, where each KV block corrects for 1 send GPS data. Another limitation is that the
out of 4 discrete samples. The remaining KV blocks position prediction is done considering only the
that have errors due to channel noise are straight forward movement and does not consider
retransmitted exactly once selectively in the next any sharp turns or obstructions. This is addressed by
ensemble since the receiver has a knowledge of the developing a possible tracking region, formed using
location of the KV blocks in error within the the joint information of possible forward movement
ensemble received. In addition, each set of discrete and sharp turns of the target, based on two previous
samples are interleaved to reduce the impact of burst positions. It is updated with the latest GPS data until
errors. It has been shown that for Eb/No of << 10 dB, the target reaches the coverage boundary. This
we can recover data at a BER of 10-7[8] in a multi- system can be employed outdoors where reasonable
path faded channel. The performance of the accuracy of GPS data is available, but this cannot be
proposed integrated zone finding and triangulation applied both indoors and indoor-outdoor moving
method is presented while minimizing the impact of radio applications due to severe multi-path
multi-path fading and other interference. interference that effectively minimizes the
We have already researched on PL&T deploying communications availability. Emergency disaster
zone forming and triangulation using two reference management applications require radios to move
nodes [1-2]. In this paper, the focus is a single both indoors and outdoors. Even in outdoor where
reference node based PL&T and comparison with large buildings exist, GPS data many not be very
prediction filter based PL&T for mobile radio accurate, thus limiting its usage.
When GPS is not accurate, the reliance to PL&T communication with the reference radio or vice-
triangulation method using neighbouring references versa. When the reference radio and the target are
to track target radios is high. Many researchers have exchanging IP packets as part of normal data
demonstrated the use of directional antennas for transfer, they are used for PL&T operation. The
increasing the coverage area and use signal PL&T control sets the number of IP packets in an
strengths and the arriving angle of the signal with an ensemble and they are time stamped at the PS sub-
established base of a pair of references whose layer. The Time of Departure (ToD) of each IP
locations are known in an adhoc network packet in the ensemble is recorded. At the remote
environment [1-3]. To improve the accuracy, radio, the Time of Arrival (ToA) of each IP packet
Kalman filter can be employed in this method, in the ensemble is recorded. Because packets get
similar to the one used in GPS based algorithm to random delays, any packet that arrives after the
improve the accuracy of PL&T measurement completion of the ensemble time is not account for
[4],[7],[9].This algorithm limits its use for motion of ToA measurement. The ToAs of packets received
the radio with limited directional change. A within the ensemble time are transmitted in a
Minimal Contour Tracking Algorithm (MCTA) is management packet to the sender. The sender
employed to concentrate tracking area where the computes the range based on the difference between
target vehicle most frequently appears [5]. Although ToA and ToD for each packet and the average of all
MCTA saves power consumption from the sensor the differences provide the range in particular
communication, the mapping process for tracking direction which is the distance as indexed to
the target with sensors is computationally complex propagation time. Specifically, directional antenna
and does not work in high speed vehicle. The energy mode is used by transmitter and omni directional
contour formed by target vehicle can be interrupted antenna as well as directional antenna mode by
or overlapped immediately by other vehicles, which receiver to compute ToA, ToD and AoA. This
contradicts the MCTA. strategy is more efficient in realtime battlefield,
when the directional change of node is in the range
2. Problem & Proposed Solution of -30 to +30 degree until there are twists and turns.
Additionally, it provides higher security as the
prediction and tracking is based on a single hop only
The problem is to formulate robust PL&T scheme
in particular directional range.
using predictive Kalman filter and zone finding
In this method, using the directional antenna with a
approach with triangulation. These PL&T schemes
beam width of theta degrees is used to find the range
need to be compared under different circumstances.
from two references to a target radio using
The bit error performance of zone finding method
triangulation. The directional antenna is moved in
needs to be optimized using OFDM system under
each reference unit, until the target radio is in the
severe multipath fading for indoor environment.
beam width of each reference unit and can communi
The proposed solution includes the following
aspects:
 Specifying and executing a Baseline method
based PL&T using predictive Kalman filter
and triangulation
 Specifying and executing a single reference
node based novel PL&T using zone finding
and triangulation
 Comparison of different PL&T schemes
 Performance analysis of novel PL&T using
interleaving KV transform coding in OFDM
based system under multipath fading
2.1 Baseline PL&T using Directional Beam with
Predictive Kalman Filter
The PL&T requires cross-layer management to Fig. 1 Radio Architecture for PL&T Computation
support between IP Layer and Physical Medium
Dependent (PMD) Layer as shown in Fig.1. At the -cate with the two references. Then the range and
IP layer, IP Packets are generated for exchange the corresponding [X,Y] co-ordinate of the target
between two radios. For PL&T operation, the IP radio is computed to identify the initial location. An
packets are generated by the PL&T control function extended Kalman Filter is used for recursive
in the management when the target is not in prediction, computation and correction of the future
PL&T of the target radio over time based on the G k+1 =Pk+1/k HT k+1 S-1k+1 (7)
limited directional path of the target radio and its
availability in both beams of the references. Once
iii) Corrected error covariance, true state position
the target radio is outside one of the beam, then a
and recursively continue to (i), (ii) and (iii).
decision is taken that it is out of range and two new
reference nodes are recruited.
The co-ordinates of reference node R(xi,yi) which is Pk+1 =Pk+1/k -G k+1Hk+1Pk+1/k (8)
taken as (0,0) and co-ordinates of transmitting
node S(xj,yj) which is taken as (x, y) depending For unconstrained Kalman Filter;
upon the received signal strength as shown in Fig. 2.
Then, Angle of Arrival (AoA) of signal from R(α) Xk 1/k  Xk 1  Gk 1 * (Yk 1  Hk 1 * Xk 1 ) (9)
and AoA of signal from S(β) are computed. By
triangulation method, coordinates of the desired For constrained Kalman Filter;
node P(xk,yk) which is being tracked can be
computed from the general equation of line PR and Xk+1/ k+1 = Xk+1/k - Pk+1/k * DTk * Inv (Dk *Pk+1/k *DTk ) * Dk * Xk+1/k
PS for the initialization as follows and can be used
(10)
as reference point for other unknown node in multi-
hop scenario as shown in Fig. 2 and equations (1)
Hence, Xk represents the initial state, Fk is the state
and (2).
transition model, T is the length of the tracking
update time interval, Gk is Kalman gain and Pk
refers to the estimation error covariance. Similarly,
Yk is the measurement observation, Uk is the control
input, Hk is the measurement matrix, Qk is the
process noise covariance, Rk is the measurement
noise covariance, Dk is the state constraint matrix,
Bk is the input matrix and θ is the heading angle.
The receiving node’s vehicle dynamics and
measurements can be initialized as follows:
Fig. 2 Illustration of co-ordinate computation

x k =((y j -yi )+xi Tanα-x jTanβ)/(Tanα-Tanβ)   (1)

Bk= , Fk=
yk =((xi -x j ) TanαTanβ+(y jTanα- yiTanβ))/(Tanα-Tanβ)   (2)

Once the initial location is achieved, the future

position is derived by the Kalman filter based Hk+1 = ,
tracking method which recursively performs the
position prediction, estimation and tracking. The
Dk=
discrete Kalman Filter can be modelled as extended
Kalman filter since the dynamic equations are linear
but the measurement equations are nonlinear to Thus, Kalman filter dynamically predict the
estimate the state vector [4],[7],[9]. covariance in estimated position in certain direction
i) Prediction of state vector, measurement with some speed, compute the Kalman gain
observation vector and error covariance are depening upon actual measurements and provides
computed as follows: the corrected true state postion and covariance
which is used to predict position in recursive order.
Xk+1= Fk Xk + Bk Uk (3) In the case of constrained Kalman filter, state
constrained is deployed so that position estimate
Yk 1  Hk 1Xk 1  R k 1 errror can be reduced as compared to the state
(4)
Pk 1/k  Fk Pk/k Fk  Qk
T
(5) unconstrained filter case. In some circumstances,
once the reference node goes out of radio range
from the transmitter then the next neighbouring
ii) Computation of Kalman Filter Gain, node is recruited as the reference node and the
Sk+1 = Hk+1 Pk+1/k HT k+1 + R k+1 (6) algorithm needs to work from the beginning.
2.2 Integrated Zone based PL&T
The integrated zone is formed considering the iv) Again, a new circle is drawn with radius rj =di/2
intersection of zone for forward movement and zone at the point B (xi,yi) and the point P(x,y) where the
for sharp turns. This method has used only one line joining the centers of two circles meets the line
reference where as the modified PL&T has used two containing the points of intersection of the two
reference nodes as well as mapping the received circles is computed as follows:
energy into radii distance in the reference paper [1]. Equation of first circle with radius ri and centre
The position prediction and tracking of desired (xi+1,yi+1) is given by:
receiver node can be done by predicting the tracking (x – xi1 )2  (y – yi1 )2  ri 2 (16)
zone which is created with some geometrical
applications as illustrated in Fig. 3. The tracking
Equation of second circle with radius rj and centre
zone is formed considering the last two positions of
(xi,yi) is given by:
the desired node and includes the joint information
of forward movement and sharp turns for ( x  xi )2  ( y  yi )2  rj 2 (17)
obstructions. This tactics also needs directional
antennas and omni directional antenna for each node The point P(x,y) where the line joining the centers
and saves drastic amount of power. Directional of two circles meets the line containing the points
Antenna mode is used by transmitter and directional of intersection of the two circles is computed.
antenna as well as directional antenna mode by
receiver [6]. This algorithm is most efficient as it x  0.5*(xi1  xi )  (xi1  xi ) (r 2i  r 2 j ) / 0.5*{(xi1  xi )2  (yi1  yi )2}
provides the tracking zone even the directional
change of node is made more than range of -30 to (18)
+30 degree for twists and turns on obstructions in y  0.5*(yi1  yi )  (yi1  yi ) (r i  r j ) / 0.5*{(x i1  x i )  (yi1  yi )2}
2 2 2

the battlefield. It is also more precise and secured as

(19)
the prediction and tracking based on single hop
achieved from the computation of arrival time. The
Then, draw a circle at point P (x,y) with radius ri
algorithm works as follows:
where the future location of neighbor node is
i) When the last two locations A(xi-1,yi-1) and B predicted and beam width must be defined.
(xi,yi) of a receiving node is achieved through CTS
signal with GPS data then a line is sketched passing v) Compute neededthe beam width
through known A and B. The formulation of this line   2 Arcsin (ri / D j ) where, ri is radius of circle
is y = mx+ c, where m is the slope and c is the y- and Dj is the distance from transmitter to P(x,y) the
intercept are calculated by following equation:
point of intersected lines. The real-time beamwidth
deployed in Figure-2 is given as:
m  (yi  yi 1 ) / (x i  x i 1 ) (11)
c  (x i yi 1  xi 1 yi ) / (x i  x i 1 ) (12) Beamwidth  1if   1;
i if i 1    i;
ii) An equilateral triangle can be formed such that
one vertex is at B position and its center (xi+1, yi+1) is n if   n
on line l using following formula and the length of
edges of this equilateral triangle is di which is the
distance between last two locations A and B.

xi1  xi  { di *(xi  xi1 )* Cos (arctan ( k )) / xi  xi1 }

(13)
yi1  yi  { di *(yi  yi1)* Sin (arctan ( k )) / 2 yi  yi1 }
(14)

iii) A circle is drawn that can include the equilateral

triangle with smallest radius such that the circle and
the equilateral triangle have the same center
(xi+1,yi+1) and the radius of the circle is ri; Fig. 3 Adaptive Beam forming over integrated zone
ri  di /  3 (15)
Once the desired node moves close to the boundary
of tracking zone then the algorithm must be repeated
depending upon last two positions achieved from
GPS data.

3. Simulation and Performance

Evaluation
The simulation is done in C and Matlab for
Baseline PL&T and Integrated Zone based PL&T
for GPS free systems respectively in 500 X 500 sq. Fig. 4 True position in the North & East directional
m area. For Baseline PL&T, a transmitter node first position.
selects the neighbouring node as reference node and
locates the receiver node which is to be tracked by
simply triangle formation with the ToA and ToD
computation and angular information. Then,
receiver node’s future position is predicted in its
northern and eastern positions as the battlefield
mobility in forward direction is +30 degree to 30
degree. The dynamic equations are linear but the
measurement equations are nonlinear, so the
extended Kalman filter is deployed to estimate the
state vector. Inside the beam coverage within certain
range after locating receiver node then it is to be
tracked and the covariance of the process and
Fig. 5 Unconstrained Kalman Filter position error
measurement noise are set as:
in the North position and East position.
Q(k)= Diag[4m/s,4m/s,1m/ s2,1m/ s2],
R(k)= Diag[900 m2,900m2]

When the receiver node vehicle is travelling off-

road, or on an unknown road, terrains then the state
position prediction is complex and this problem is
unconstrained. Rest of the case, it may be known
that the vehicle is travelling on a given road, which
is known as the constrained state estimation using
G=Inverse (P). When the vehicle is travelling on a
road with a heading angle 60 degree, sample period
T is 3 sec and preferred acceleration is set to 1 m/s2. Fig. 6 Constrained Kalman Fillter position error in
The initial conditions for state vector and error the North position and East position
variance are set as expressed in X(k)and P(k/k).

X(k)=[0 0 17 10] T, P(k/k)=Diag[900 900 4 4]T

The true position profile of receiver node is tracked

in North and East direction as shown in Figure-3
and position estimation errors are calculated and
simulated in both directional positions using
constrained and unconstrained Kalman filter as in
Fig. 4-8. The simulation results show that the
position estimation error is found lower in
constrained filter rather than unconstrained because Fig. 7 Unconstrained and constrained KF in the
the initial state is in always favoured as moving North position estimation error
along the predefined road in the constrained filter.
The average position error is about 5 m in the North
position and 3m in East position for unconstrained
filter whereas about 1m in both the North and East
positions for constrained filter.

Fig. 10 Efficiency versus beam width

Additionally, the efficiency is found inversely

proportional to the silent period. In battlefield a
vehicle moving with average velocity of 60km/h,
Fig. 8 Unconstrained and constrained KF in the East can covers more than 330 meters during 20 seconds
position estimation error. or 500 meters during 30 seconds and can change its
direction with an absolute relative direction angle
The algorithm of position prediction and tracking being lager than 30 degree. The efficiency decreases
with GPS is simulated with beam width 10o-20o and to 0.38 with increasing the silent period to 30
15 seconds silent duration in 10 different Seconds. Again, the proposed algorithm has higher
experiments taking 10 Bernoulli trails at different efficiency against silent period as compared with the
position inside the predicted tracking zone on each existing algorithm as shown in Fig. 11.
experiment. The results show that the average
efficiency of the proposed algorithm is
approximately 99% whereas the existing algorithm
has 96% as shown in Fig. 9.

Fig. 11 Efficiency versus silent duration

3.1 Performance of Integrated Zone based

PL&T in OFD M system using
Fig. 9 Efficiency comparison of Proposed and Interleaving-KV Transform Coding
Existing Algorithm Integrated Zone based PL&T is deployed in OFDM
(Orthogonal Frequency Division Multiplexing)
Similarly, while simulating the efficiency with beam based IEEE 802.11(a) system specification as
width, the efficiency is found directly proportional illustrated in Table-1 for robust performance in the
to beam width for the battlefield characteristics of dispersive channels. The OFDM based system is
directional change about 30 degree. The efficiency deployed at the physical layer with Integrated Zone
of 0.7 is found at 2 degree because the coverage based PL&T application. The OFDM parameters for
area is very small in such case. The efficiency is simulation are listed in the Table-1, to analyze the
0.95 at the beam width between 7 to 14 and 1 Bit Error Rate (BER) performance against Eb/No (bit
between 28 and 30. The proposed algorithm energy/noise). First of all, transmitter runs the
(Integrated Zone based PL&T) has higher efficiency PL&T algorithm at application layer to localize and
against beam width as compared with the existing track the desired receiver sending KV blocks stream
algorithm (LAPDC) as shown in Fig. 10. to the receiver after handshaking. KV transform
coding is deployed at the physical layer with rest 12 subcarriers are wasted to roll up the
OFDMA to enable the radio channel to handle spectrum. For this scenario, regarding the available
multi-path fading by recovering the data with low bandwidth from -10MHz to +10MHz, only
Bit Error Rate (BER) at low SNR or Eb/No. The subcarriers from -8.1250MHz (-26/64*20MHz) to
transmitter side of the KV transform coding convert +8.1250MHz (+26/64*20MHz) are used. In other
a set of bits to a discrete sample. Each KV transform words, the signal energy is spread over a bandwidth
uses four discrete samples to produce four of 16.25MHz but the noise is spread over bandwidth
coefficient samples at the output and two overhead
samples for error correction. By modifying these Table 1: IEEE 802.11a Specification Parameters
real time samples, the transmission samples will
have finite voltages that can be coded by discrete Parameter Value
codes. The number of bits per transmission sample Modulation BPSK
is same as the number of bits/sample at the input. FFT size 64
The transmitter side uses multiple KV blocks each No. of used Subcarrier (nDSC) 52
producing the transmission samples and these FFT sampling frequency 20 MHz
samples are interleaved in packets. Then, these Subcarrier spacing 312KHz
packets are sent through OFDM based transmitter, Used subcarrier index {-26 to -1, +1 to +26}
channel (Rayleigh and Rician) and OFDM receiver. Cyclic prefix duration (Tcp) 0.8μs
The KV transform at the decoder receives the Data Symbol duration (Td) 3.2μs
estimated coefficient samples after sample Total Symbol duration(Ts) 4μs
correction and then the bits are recovered. The KV
system is implemented to include the transmission of 16.250MHz, whereas noise is spread over
of ensembles of packets, single sample error bandwidth of 20MHz. The cyclic prefix is guard on
correction in each KV block, sample interleaving 16 samples as 0.8 μs from the end of the sinusoidal
from a set of KV blocks, the single retransmission appended to the beginning of the sinusoidal to
of selected KV blocks in error in each ensemble. mitigate the natural time dispersion among symbols.
There are two different multipath channel model This prevents signal discontinuities and achieves the
deployed separately known as Rayleigh channel in original sinusoidal of frequency 312.5 KHz.
the absence of line of sight and Rician channel in
the presence of line of sight, and then OFDM at the 0
PL&T Performance in Rayleigh Channel
10
physical layer. The Rayleigh fading channel models
that the magnitude of a signal that passed through a -1
10

transmission medium will vary randomly, according

-2
to the radial component of the sum of two 10
Bit Error Rate

uncorrelated Gaussian random variables. The -3

10
Rayleigh channel is model as 10 tap channel such
that the real and imaginary part of each tap is an -4
10

independent Gaussian random variable. On the other -5

BPSK/OFDM
10 KV/BPSK/OFDM
hand, The Rician channel has dominant line of sight KV-Interleaving/BPSK/OFDM

in addition to Rayleigh distribution. The best and -6

10
0 2 4 6 8 10 12 14 16
worst scenario of Rician fading channels depends Eb/No (db)

upon k-factors, k=β2/2σ2 where β is the amplitude of

the spectacular component σ is the variance of zero Fig. 12 PL&T performance in Rayleigh Channel
mean stationary Gaussian process. The Rician
channel with K = ∞, is the Gaussian channel with PL&T is deployed by encoding PL&T data
strong line of sight whereas the Rician channel with (distance and direction) bits with three different
K = 0, is Rayleigh channel with no line of sight modulation schemes categorized as BPSK/OFDM,
path, respectively. The Rician channel with k=1 and KV/BPSK/OFDM and KV/Interleaving/BPSK/
random noise is deployed in simulation. OFDM in both Rayleigh and Rician channel. The
For multicarrier modulation, the symbol duration is BER is found inversely proportional to the Eb/No or
3.2 μs for subcarriers space +- 312.5KHz, +- transmit power in both channels. In other words, the
625KHz… and so far. The available bandwidth of BER decreases as the Eb/No increases and then
20MHz is split into 64 subcarriers. But, out of the optimizes at some point due to burst errors as shown
available 64 subcarriers, only 52 subcarriers are in Fig. 12-13. This refers that bit stream can be
used for transmitting the sequence of KV blocks and recovered perfectly in higher Eb/No until the
deleterious burst errors are appeared and after that unconstrained scenario for dynamic tracking. The
BER cannot be increased even Eb/No increases. position estimation errors are found lower in
constrained filter in the North and East position
PL&T Performance in Rician Channel
0
10 estimation because there is initial state constrained
BPSK/OFDM
-1 KV/BPSK/OFDM
for smoothing the covariance errors. Similarly,
10
KV-Interleaving/BPSK/OFDM
Integrated Zone based PL&T (IZPL&T) algorithm
-2
10 creating tracking zone, deploying the joint
information of forward movement and sharp turns,
Bit Error Rate

-3
10
has better performance than LPADC. Both the
-4
10 devised algorithm considers the battlefield mobility
-5
10
characteristics, spatial reuse, low power
consumption, security and precision. The simulation
-6
10
illustrates the outstanding performance of IZPL&T
-7
10 with KV/Interleaving/BPSK/OFDM modulation
0 2 4 6 8 10 12 14 16
Eb/No (db) scheme for both Rician and Rayleigh channel.
Future research will concentrate on more secured
Fig. 13 PL&T performance in Rician Channel multiple targets’ PL&T with effective zonal
computation in mobile ad hoc network fabrics.

From the simulation results, BER is found ACKNOWLEDGMENT

drastically reduced in This research work is supported in part by the U.S.
KV/Interleaving/BPSK/OFDM modulation scheme ARO under Cooperative Agreement W911NF-04-2-
in both Rayleigh and Rician channel. Therefore, 0054 and the National Science Foundation NSF
KV/Interleaving/BPSK/OFDM outperforms over 0931679. The views and conclusions contained in
KV/ BPSK/OFDM and BPSK/OFDM, while KV/ this document are those of the authors and should
BPSK/OFDM outperform over BPSK/OFDM in not be interpreted as representing the official
both Rayleigh and Rician channel as illustrated by policies, either expressed or implied, of the Army
Fig. 12-13. The major reason is that KV with Research Office or the National Science Foundation
interleaving in BPSK over OFDM can efficiently or the U. S. Government.
recover the data transmitted at receiver as it
deployed the correlation between data sets during 5. References
KV encoding and KV decoding. Similarly, KV in
BPSK over OFDM can slightly recover data [1] Shakhakarmi, N., Dhadesugoor, R.,V.: Distrib-
transmitted at receiver using simple error correction -uted Position Localization and Tracking
and data retransmission. Furthermore, the Eb/No is (DPLT) of Malicious Nodes in Cluster Based
found lower in Rician channel as compared to Mobile Adhoc Networks (MANET), WSEAS
Rayleigh channel by 1 db, 2 db, 2 db,2 db, 2 db at Transactions in Communications, ISSN: 1109-
BER of 10-2, 10-3, 10-4, 10-5, 10-6 for 2742, Issue 11, Volume 9, November 2010.
KV/Interleaving/BPSK/OFDM scheme. In other [2] Shakhakarmi, N., Dhadesugoor, R.V.: Dynamic
words, the higher BER is found in Rayleigh channel PL&T using Two Reference Nodes Equipped
rather than Rician channel due to the higher carrier with Steered Directional Antenna for Significant
frequency offset and delay spread in the absence of PL&T Accuracy, Wireless Telecommunications
line of sight which violate the orthogonality among Symposium (WTS 2012), Lond on, UK, April
sub carriers and increase the Inter Carrier 18-20, 2012.
Interference (ICI). [3] Roy, S., Chatterjee, S., Bandyopadhyay, S.,
Ueda, T., Iwai, H., Obana, S.: Neighborhood
4. Conclusion Tracking and Location Estimation of Nodes in
Ad hoc Networks Using Directional Antenna: A
In MANET, a receiver node’s real time position is Test bed Implementation, Proceedings of the
predicted, estimated and tracked by a transmitter Wireless Communications Conference, Maui,
when each node has directional antenna. Real time Hawaii, USA, June 13-16, 2005.
means the location is determined inside the tracking [4] Simon, J., Jeffrey, J., Uhlmann, K.: A New
zone before mobile node movement. In GPS free Extension of the Kalman Filter to Nonlinear
system, Baseline PL&T is deployed with extended Systems, The Robotics Research Group,
Kalman filter considering the state constrained and Department of Engineering Science, The
 University of Oxford, UK. SAE, WTS and editorial board member for IJEECE,
[5] Jeong, J., Hwang, T., He, T., Du, D.: MCTA: WASET, JCS and IJCN journals.
Target Tracking Algorithm based on Minimal
Contour in Wire Sensor Networks, In proc.
IEEE Infocom, 2007. Dhadesugoor R. Vaman is
[6] Lu, X., Wicker, F., Leung, I., Li`o, P., Xiong, Z.: Texas A & M University
Location Prediction Algorithm for Directional Board of Regents and Texas
Communication, Computer Laboratory, Instrument Endowed Chair
University of Cambridge, U.K., and Beijing Professor and Founding
University of Aeronautics and Astronautics , Director of ARO Center for
China, IEEE, 2008. Battlefield Communications
[7] Welch, G., Bishop, G.: An Introduction to the (CeBCom) Research, ECE
Kalman Filter, Department of Computer Department, Prairie View A&M University
Science, University of North Carolina, July 24, (PVAMU). He has more than 38 years of research
2006. experience in telecommunications and networking
[8] Dhadesugoor, R.,V., Koay, S., Annamalai, A., area. Currently, he has been working on the control
Agarwal, N.: A Simple and Least Complex KV based mobile ad hoc and sensor networks with
(Koay-Vaman) Transform Coding Technique emphasis on achieving bandwidth efficiency using
with Low BER Performance at Low Eb/N0 for KV transform coding; integrated power control,
Multi-Tiered Applications in Power and scheduling and routing in cluster based network
Bandwidth Constrained MANET / Sensor architecture; QoS assurance for multi-service
Networks, Proceedings of IEEE SMC 2009, applications; and efficient network management.
October 2009. Prior to joining PVAMU, Dr. Vaman was the CEO
[9] Menglong, C., Lei, Y., Cui: Simultaneous of Megaxess (now restructured as MXC) which
Localization and Map Building Using developed a business ISP product to offer
Constrained State Estimate Algorithm, Institute differentiated QoS assured multi-services with
of Autonomous Navigation and Intelligent dynamic bandwidth management and successfully
Control, Chinese Control Conference, Qingdao, deployed in several ISPs. Prior to being a CEO, Dr.
China, 2008. Vaman was a Professor of EECS and founding
Director of Advanced Telecommunications
Niraj Shakhakarmi worked Institute, Stevens Institute of Technology (1984
as a Doctoral Researcher since 1998); Member, Technology Staff in COMSAT
2009-2011 in the ARO Center (Currently Lockheed Martin) Laboratories (1981-
for Digital Battlefield 84) and Network Analysis Corporation (CONTEL)
Communications (CeBCom) (1979-81); Research Associate in Communications
Research, Department of Laboratory, The City College of New York (1974-
Electrical and Computer 79); and Systems Engineer in Space Applications
Engineering, Prairie View A&M University. He Center (Indian Space Research Organization) (1971-
received his B.E. degree in Computer Engineering 1974). He was also the Chairman of IEEE 802.9
in 2005 and M.Sc. in Information and ISLAN Standards Committee and made numerous
Communications Engineering in 2007. His research technical contributions and produced 4 standards.
interests are in the areas of Wireless Dr. Vaman has published over 200 papers in
Communications and Networks Security, Secured journals and conferences; widely lectured nationally
Position Location & Tracking (PL&T) of Malicious and internationally; has been a key note speaker in
Radios, Cognitive Radio Networks, many IEEE and other conferences, and industry
WCDMA/HSPA/LTE/WRAN, Next Generations forums. He has received numerous awards and
Wireless Networks, Satellite Networks and Digital patents, and many of his innovations have been
Signal Processing, Wavelets Applications and successfully transferred to industry for developing
Image/Colour Technology. He is a member of IEEE commercial products.
Communications Society, ISOC, IAENG and
attended AMIE conference. He has published
several WSEAS and IJCSI journals, WTS, Elsevier
and ICSST conference paper. His several journals
and conference papers are under review in IEEE
journals. He is serving as reviewer for WSEAS,