AN EFFICIENT IMPLEMENTATION OF TRACKING USING KALMAN FILTER FOR UNDERWATER ROBOT APPLICATION

International Journal of Computer Science, Engineering and Information Technology (IJCSEIT), Vol.2, No.2, April 2012
DOI : 10.5121/ijcseit.2012.2207 67
AN EFFICIENT IMPLEMENTATION OF TRACKING
USING KALMAN FILTER FOR UNDERWATER
ROBOT APPLICATION
Nagamani Modalavalasa1
, G SasiBhushana Rao2
, K. Satya Prasad3
1
Dept.of ECE, SBTET, Andhra Pradesh, INDIA
mani.modalavalasa@gmail.com
2
Dept. of ECE, Andhra University, Visakhapatnam, Andhra Pradesh, INDIA
3
Dept.of ECE, Jawaharlal Nehru Technological University Kakinada, Kakinada, INDIA
ABSTRACT
The exploration of oceans and sea beds is being made increasingly possible through the development of
Autonomous Underwater Vehicles (AUVs). This is an activity that concerns the marine community and it
must confront the existence of notable challenges. However, an automatic detecting and tracking system is
the first and foremost element for an AUV or an aqueous surveillance network. In this paper a method of
Kalman filter was presented to solve the problems of objects track in sonar images. Region of object was
extracted by threshold segment and morphology process, and the features of invariant moment and area
were analysed. Results show that the method presented has the advantages of good robustness, high
accuracy and real-time characteristic, and it is efficient in underwater target track based on sonar images
and also suited for the purpose of Obstacle avoidance for the AUV to operate in the constrained
underwater environment.
KEYWORDS
Autonomous Underwater Vehicle, Tracking, SONAR, Threshold, Obstacle avoidance
1. INTRODUCTION
Autonomous underwater vehicles (AUVs) have the potential to revolutionize our access to the
oceans to address critical problems facing the marine community such as underwater search and
mapping, climate change assessment, marine habitat monitoring, and shallow water mine
countermeasures. Navigation is one of the primary challenges in AUV research today.
Navigation is an important requirement for any type of mobile robot, but this is especially true for
autonomous underwater vehicles. Good navigation information is essential for safe operation and
recovery of an AUV. For the data gathered by an AUV to be of value, the location from which the
data has been acquired must be accurately known. Some of the important concerns for AUV
navigation, such as the effects of acoustic propagation are unique to the ocean environment.
The goal of this paper is the Surveillance using Imaging Sonar Data for Underwater Robot
Application based on Kalman filter. In this paper, the images from the Compressed High Intensity
Radar Pulse (CHIRP) Sonar are used for the analysis. CHIRP Sonars are the active Sonars
invented to overcome the limitations of conventional monotonic Sonars. In conventional Sonars
when the separation of the targets is less than the range resolution, then it displays as a single

68
large combined target rather than the multiple smaller targets. On the other hand, if we use the
smaller transmission pulse to increase the range resolution, then the maximum range obtained
decreases due to less energy. In order to overcome this problem, CHIRP Sonars have been
developed and made use of. . In this paper Sector Scan SONAR images are taken as input data
and processed for obtaining the tracking results.
Several methods are available for tracking the objects in the image sequences received from the
Sonar fitted on the AUV. These methods are not suitable for undertaking the collision avoidance
if the AUV is required to be controlled in the constrained underwater environment. The time
required for extraction of target parameters and for subsequent tracking becomes an important
criterion because the AUV is required to be maneuvered well before the collision occurs. A
comparison of the commonly used algorithms for data association and tracking namely Nearest
Neighbour Kalman filter (NNKF) and Probabilistic data association filter (PDAF) is made in ref.
[1] for single target tracking in clutter. In this paper, tracking of the objects (detected in the
sequence of images received from Sonar) based on their centroids has been presented.
Accordingly the calculation of the centroid, tracking of the objects and the calculation of the
trajectories has been presented.
2. IMAGE PROCESSING MODEL BASED ON CENTROID OF THE OBJECT
In image-based air traffic control or air defense system, automatic detection and tracking of
targets are extremely important for their safety or early warning. In such scenario, the sensor
images are often cluttered, dim, spurious or noisy due to the fact that the distances to targets from
the control centre are large. Tracking problems involve processing measurements from a target of
interest and producing at each time step, an estimate of the target’s current position and velocity
vectors. Uncertainties in the target motion and in the measured values, usually modelled as
additive random noise, lead to corresponding uncertainties in the target state. Also, there is
additional uncertainty regarding the origin of the received data, which may or may not include
measurements from the targets and may be due to random clutter (false alarms). This leads to the
problem of data association [2]. In this situation tracking algorithms have to include information
on detection and false alarm probabilities. This approach provides a method of centroid tracking
and target identity estimation using image SONAR data.
The Centroid tracking combines both object and motion recognition characteristics for practical
target tracking from imaging sensors. The characteristics of the image considered are the intensity
and size of the cluster. The pixel intensity is discretised into several layers of gray level intensities
and it is assumed that sufficient target pixel intensities are within the limits of certain target
layers. The centroid tracking implementation involves the conversion of the image into a binary
image and applying upper and lower threshold limits for the “target layers”. The binary target
image is then converted to clusters by using nearest neighbour criterion. If the target size is
known, then it is used to set limits for removing those clusters that differ sufficiently from the
size of the target cluster to reduce computational complexity. The centroid of the clusters is then
calculated and this information is used for tracking the target. The Centroid tracking involves the
following steps:
a. Pre-processing to remove the noise / blur from the images. (In present day
applications this step is generally performed by the Sonar)
b. Identifying potential targets by image segmentation methods. Real Sonar image
and the image after segmentation are shown in Figure 1 and Figure 2. In this,
the image is segmented into objects, shadow and sea bottom reverberation
regions and then the edges of the object are extracted [3].
c. Calculation of the centroids for all the detected objects.

69
d. The steps (a) to (c) are performed on all subsequent images
e. Identification of the moving and stationary objects.
f. Determination of the association of the moving objects based on the maximum
speed criterion.
g. Tracking of the moving objects using Kalman Filter [4, 5]
h. Calculation of the trajectory.
i. Calculation of Collision course by taking the own speed and direction into
consideration [6].
j. Finally executing the manoeuvring commands to AUV
3. TRACKING ALGORITHM
The sequence of images can either be processed in real-time, coming directly from a video
camera for example, or it can be performed on a recorded set of images. The implementation of
this paper uses recorded image sequences although the theory can be applied to both types of
applications. The target to be tracked might be a complete object or a small area on an object. In
either case, the feature of interest is typically contained within a target region. In this paper will
consider target centroid positions across the image plane. The position will be described in X-Y
coordinates in pixel units on the image ( i.e. image coordinates).
Figure 2. Object identification from
the proposed segmentation method
Figure 1. Real SONAR image

70
Once objects are detected, we extract their outline and track their centroids across the image plane
using separate linear Kalman filters to estimate their x and y coordinates. The Kalman filter
provides a general solution to the recursive minimised mean square linear estimation problem.
The mean square error will be minimised as long as the target dynamics and the measurement
noise are accurately modelled. In addition, the Kalman filter provides a convenient measure of the
estimation accuracy through the covariance matrix, and the gain sequence automatically adapts to
the variability of the data. Figure 3 explains the steps, those have been implemented to track the
objects in the images received from the imaging Sonar of AUV. A linear, discrete-time dynamic
system describing the target parameters estimation is shown in Figure 4.
4. KALMAN FILTER
Derivations of the Kalman filter are available in the literature, e.g. [7], [8] and [9]. The transition
from one state to the next could be described in many ways. These different alternatives can be
grouped into linear and non-linear functions describing the state transition. Although it is possible
to handle either of these transition types, the standard Kalman filter employs a linear transition
function. The extended Kalman filter (EKF) allows a non-linear
transition, together with a non-linearmeasurement relationship. For the standard Kalman filter, the
state transition from k-1 to k can be expressed with the equation
xk = Axk-1 + wk-1 (1)
where A is referred to as the state transition matrix and w k-1 is a noise term. This noise term is a
Gaussian random variable with zero mean and a covariance matrix Q, so its probability
distribution is
Calculation of object
centroid in the image
Comparison of centroid
values in all subsequent
frames
Identification of moving
objects based on change
in centroid values
Separation of moving
objects from the
stationary objects
Calculation of speed and
direction of moving
objects
Calculation of trajectory
for all moving objects
Fig.3 Block diagram of the complete model
Fig. 4 Target aspects estimation and tracking from Kalman filter
Z-1 xk Hk∑ ∑
Ak+1, k
vk
wk zk
Process equation Measurement equation

71
p(w) ~ N(0,Q) (2)
The covariance matrix Q will be referred to as the process noise covariance matrix in the
remainder of this report. It accounts for possible changes in the process between k-1 and k that
are not already accounted for in the state transition matrix. Another assumed property of wk-1 is
that it is independent of the state xk-1 .
It is also necessary to model the measurement process, or the relationship between the state and
the measurement. In a general sense, it is not always possible to observe the process directly (i.e.
all the state parameters are observable without error). Some of the parameters describing the state
may not be observable at all, measurements might be scaled parameters, or possibly a
combination of multiple parameters. Again, the assumption is made that the relationship is linear.
So the measurement zk can be expressed in terms of the state xk [10] with
zk = Hxk + vk (3)
where H is an m × n matrix which relates the state to the measurement. Much like w k-1 for the
process, vk-1 is the noise of the measurement. It is also assumed to have a normal distribution
expressed by
p(v) ~ N(0,R) (4)
where R is the covariance matrix referred to as measurement noise covariance matrix.
In our analysis, the state xk contains the position (x, y) of the object at the instant k and also the
speed of the object in both x ( x& ) and y ( y& ) directions [11]. The new position (xk, yk) is the old
position (xk-1, yk-1) plus the velocity ( 1−kx& , 1−ky& ) plus noise w k-1.
The state equation, Eq. (5) is then in this case is defined as
1
1
1
1
1
1000
0100
010
001
−
−
−
−
−
+
























=












k
k
k
k
k
k
k
k
k
w
y
x
y
x
t
t
y
x
y
x
&
&
&
&
(5)
Where ‘t’ represents the time interval between any two successive image frames and is
considered as 1second. And the measurement equation Eq. (6) is written as
k
k
k
k
k
kmeas
kmeas
k v
y
x
y
x
y
x
z +


















=





=
&
&0010
0001
(6)
Where xkmeas and ykmeas are the measured positions in x and y directions.
The Kalman filter estimates a process by using a form of feedback control: the filter estimates the
process state at some time and then obtains feedback in the form of (noisy) measurements. As

72
such, the equations for the Kalman filter fall into two groups: time update equations and
measurement update equations. The time update equations are responsible for projecting forward
(in time) the current state and error covariance estimates to obtain the priori estimates for the next
time step. The measurement update equations are responsible for the feedback
i.e. for incorporating a new measurement into the priori estimate to obtain an improved
posteriori estimate.
Figure 5. Kalman filter cycle
As shown in Figure 5, the time update projects the current state estimate ahead in time and the
measurement update adjusts the projected estimate by an actual measurement at the time.
The time update equations can also be thought of as predictor equations, while the measurement
update equations can be thought of as corrector equations.
The main steps of Kalman Filtering algorithm that has been implemented in this paper are as
follows:-
Time Update ( Predict ) equations:
Step 1: Project the state ahead:
kkk wxAx += −
−
1
ˆˆ (7)
Step 2: Project the error covariance ahead:
QAAPP T
kk += −
−
1 (8)
Measurement Update ( Correct ) equations:
Step 3: Compute the Kalman gain:
1
)( −−−
+= RHHPHPK T
k
T
kk (9)
Step 4: Update estimation with measurements:
)ˆ(ˆˆ −−
−+= kkkkk xHzKxx (10)
Step 5: Update the error covariance:
−
−= kkk PHKIP )(
(11)
Step 6: Go to Step 1.
Time Update
( “Predict” )
Measurement Update
( “Predict” )

73
Steps 1 and 2 are responsible for projecting forward (in time) the current state and error
covariance (
−
kP ) estimates to obtain the a priori (
−
kxˆ ) estimates for the next time step. Steps 3
to 5 are responsible for the feedback i.e. for incorporating a new measurement into the a priori
estimate to obtain an improved a posteriori estimate ( kxˆ ).The Kalman gain Kk i.e. equation (9)
in (step 3) is chosen to be the gain that minimizes the posteriori error covariance. The next step is
to actually measure the process to obtain zk , and then to generate a posteriori state estimate by
incorporating the measurement as in equation (10). The final step is to obtain a posteriori error
covariance estimate via equation (11).
After each time and measurement update pair, the process is repeated with the previous posteriori
estimates used to project or predict the new priori estimates. This recursive nature is one of the
very appealing features of the Kalman filter. It makes practical implementations much more
feasible than (for example) an implementation of a Wiener filter (Brown and Hwang 1996) which
is designed to operate on all of the data directly for each estimate. The Kalman filter instead
recursively conditions the current estimate on all of the past measurements.
5. DATA ASSOCIATION
In many applications, knowledge of which target in the current frame relates to which target in
the previous frame is important and so the data association problem would need to be addressed.
Traditional multi-target tracking is based on coupling trackers such as Kalman filters, extended
Kalman filters or particle filters with a data association technique (Bar-Shalom [12] provides a
comprehensive treatment). The aim of the data association process is to interpret which
measurements are due to the targets and which are due to false alarms. An example of this used
on forward-looking sonar data is shown in [13]. Another technique which has been applied to
sonar imagery uses Optical Flow calculations to estimate direction motion [14].
Data Association plays a very important role in all tracking applications, especially in the
environment which is heterogeneous and having frequent occlusion conditions. This is aptly
applicable to the undersea environment where AUV operates.
To check that the objects those have appeared in the subsequent images actually belong to the
same target or not, following two conditions have been considered in our model [6]:-
a. The maximum pixels by which the centroid values will change in subsequent
frames, if the objects are moving with the maximum speed.
b. The variation in the position of the centroid due to the movement of water body/
occlusion.
For calculating the maximum pixels by which the centroid values will change in subsequent
frames, following model has been proposed:-
a. Assume, maximum speed that any object can have is ‘X’ km/hr.
b. The time interval between every two subsequent frames is ‘n’ seconds.
c. Image size is assumed to be kxk (for example 600x600).
d. The Range of the Sonar (R) has been taken in terms of meters.
Based on the above inputs, the maximum amount of distance (in meters) that an object will cover
when it is moving with maximum speed (X) is calculated. The distance is calculated as follows:-

74
a. Maximum distance that will be covered by the objects between consecutive
frames is [n*(X*1000/3600) ] Meters
b. The resolution of the image is R/ k meters per pixels both in row and column.
Therefore the maximum distance that will be covered in terms of pixels is given
by
[(X*1000/3600)*n]*[R/k] (12)
In this paper, following values have been selected:-
Maximum speed of the object: X = 8 km/hr.
Time difference between two subsequent frames as 1sec.
The Size of the Sonar image as 600x600.
The Sonar range (R) as 10 meters.
The resolution of the image is therefore (R/k) 10/600=0.0167 meters/pixel for both row and
column pixels.For any specified range up to 300mts, sonar can scan 0o to 360o
. While taking the
sonar data the sector is limited to 120o
and the range is kept to 10mts. Since the image size is of
600x600 where each column containing 600 pixels exactly covers 10mts range. Hence in any
direction throughout the 120o
sector of the image, the resolution of the image is 0.0167 mts/pixel.
By substituting the above values in the Eq. (12), the maximum distance in terms of meters that
will be covered by the object for every 1 second is calculated as 8.889 meters. Therefore, if the
distance between the centroids of object in the consecutive images is less than or equal to the
8.889 meters then it is assumed to be from the same object i.e. they are said to be associated.
Inherently, there is a variation in the position of the centroid due to the movement of water body.
By comparing the centroid values of the associated object in subsequent frames, if the difference
between the centroids is non-zero and is also greater than the variation in the position of the
centroid due to the movement of water body, object is said to be associated and moving. Similarly
if the centroid remains within the specified values (depends on the water body movement) in all
subsequent frames then it is assumed to be stationary. The speed of the moving object is
calculated accordingly.
Once the data association has been applied, we can initiate the tracking. Three cases are then
possible:
a. There is a new observation matching the predicted position. The Kalman filter
recursion is applied, a new state vector derived and new internal values
computed.
b. No new observation matches the prediction. The obstacle prediction is updated
using the Kalman filter internal values which are not updated. If no match is
found between the observations and a given tracked object on a predefined
number of frames, the tracked object is discarded as a false alarm.
c. An observation is not associated with any tracked object, a new object is created
and its corresponding Kalman filter initialized.
By invoking the tracking algorithm for the moving objects, their trajectories are then calculated.
6. RESULTS AND DATA PROCESSING
To acquire the required data, several experiments have been conducted at the Towing Tank of the
DRDO (NSTL), Visakhapatnam. The experiments included different scenarios such as object is
moving and Sonar remains stationary; Sonar is moving and object remains stationary; and both
Sonar and object are moving. The Towing tank is 500 meters in length with 8 meters in depth. In

75
this paper, we have implemented the proposed algorithm for the specific case i.e. object moving
and Sonar remains stationary. In this experiment, Digital Sector Scan (DSS) Sonar is fixed to
carriage and kept stationary at one place in the water, and object is moved manually towards the
Sonar. During data collection, the sector of Sonar is set to 120o
. The recorded video was then
converted into image frames using the ‘SeanetImageOut.exe’ software supplied along with the
Sonar. For the real-time implementation of the proposed algorithm, the resolution is kept low so
as to get the images instaneously at high speed.
Table 1. Comparison of Measured and Estimated Object positions (CENTROIDS)
Measured positions Estimated positions
Frames x y
coordinates
x y
coordinates
Frame1 317.3450 107.5848 317.3561 107.5907
Frame2 368.2032 187.0239 368.1796 187.0015
Frame3 473.6369 251.8636 474.2643 252.2195
Frame4 475.9537 288.6233 475.6405 288.4390
Frame5 394.5096 336.3308 393.3391 335.6495
Frame6 359.0980 398.4558 358.3820 398.0353
Frame7 370.5565 463.4487 370.2715 463.2771
The images were generated at every 1 second interval. These images were segmented so as to
identify the object and this was followed by the estimation of trajectory using Kalman Filter. In
the proposed model the objects have been associated by using the Data Association algorithm as
discussed in section 2.2. From the results the Mean Square Error for x-coordinate is found to be
0.3509 meters and for the y-coordinate is found to be 0.1188 meters. The accuracy of the object
positions is found to be ±0.0486meters. Table 1 shows the measured and estimated positions (in
terms of pixel values) of the objects in 7 subsequent frames. The final result has been shown in
Figure 5 which is a polar plot shows tracking of the object using the measured as well as the
estimated positions. From the proposed model, it is observed that time taken to identify the
objects through segmentation and extraction of obstacle parameters such as their range, bearing,
size, shape, speed and course using Kalman filter is approximately 0.4 seconds which is
reasonably less and aptly applicable for the application of obstacle avoidance. For every frame,
the complete processing takes 0.4 seconds and the time interval between the input frames is 1
second.
The sonar used to collect the data is Super Seeking DST (Digital Sonar Technology) Dual
Frequency CHIRP Sonar. Sector scan sonar with the following specifications:
Operating frequency (low)-Chirping from 250 to 350 kHz (300)
Operating frequency (high)-Chirping from 620 to720 kHz (670)
Optional high frequency 1 MHz
Beamwidth, vertical 20° [300]
Beamwidth, vertical 40° [670]
Beamwidth, horizontal 3.0° [300]

76
Beamwidth, horizontal 1.5° [670]
Maximum range 300 m [300]
Maximum range 100 m [670]
Minimum range 0.4 m
Scanned sector Variable to 360°
Range resolution 5 - 400 mm depending on range
x
y
observed
estimated
Figure 6. Object Tracking Using Kalman
7. CONCLUSIONS
Considering the fact that the time taken for calculation of target parameters for the purpose of
obstacle avoidance must be as less as possible and the position should be as accurate as possible,
the algorithm that has been proposed and validated in this paper can be concluded to meet the
given requirement. It is seen from the results that the time taken for undertaking complete
processing on every image is approximately 0.4 Seconds and the positional accuracy is found to
be ±0.0486 meters. It implies that if the obstacle is detected at the range of 300 meters and if the
AUV is moving with the speed of 8 Km/Hr then the sufficient time is available with the AUV to
take corrective course of action which is approximately 150 Seconds if the object is stationary
and approximately 75 Seconds if the object is moving head on with the same speed as AUV.
Therefore it can be concluded that the algorithm proposed in this paper is aptly suited for the
application of obstacle Avoidance in case of AUV navigating in constrained underwater scenario.
7. ACKNOWLEDGEMENTS
The above work has been undertaken towards the research project of DRDO (NSTL). The authors
are thankful to the project team at NSTL for providing the SONAR data, constant technical
support and encouragement. Authors are also thankful to the management of their respective
organizations.

77
REFERENCES
[1] Michael. J. Smith "Bayesian Sensor Fusion a framework for using multi-modal sensors to estimate
target location and identities in a battlefield scene" PhD thesis, Florida state University, 2003.
[2] Bir Bhanu, "Automatic Target Recognition: State of the Art Survey", IEEE transactions on aerospace
and electronics systems, vol. AES-22, No.4 july 1996.
[3] Yifeng Zhu and Ali Shareef, “Comparisons of Three Kalman Filter Tracking Algorithms in Sensor
Network”, this work was supported by National Science Foundation grant 0538457 and an UMaine
Startup Fund.
[4] I.T. Ruiz, Y. Petillot, D. Lane, J. Bell, “Tracking objects in underwater multibeam sonar images,”
IEE Colloquium on Motion Analysis and Tracking, London, UK, 10th May 1999, pp.11/1-11/7.
[5] I.Tena Ruiz, Y. Petillot, D. M. Lane, C. Salson “Feature Extraction and Data Association for AUV
Concurrent Mapping and Localisation” Proceedings of the 2001 IEEE International Conference on
Robotics & Automation Seoul, Korea, May 21-26, 2001.
[6] Xu, L.; Landabaso, J. L.; Lei, B.; “Segmentation and tracking of multiple moving objects for
intelligent video analysis”, BT Technology Journal, Vol 22, No 3, July 2004.
[7] S. Blackman and R. Popoli, Design and Anal- ysis of Modern Tracking Systems, Boston, MA, Artech,
pp. 157-160, 1999.
[8] A. S. Gelb, Applied Optimal Estimation, Cambridge, MA, MIT Press, 1974.
[9] Y. Bar-Shalom and T. E. Fortmann, Tracking and Data Association, Orlando, FL, Academic Press,
1988.
[10] T.H. Lee, W.-S. Ra, T.S. Yoon and J.B. Park, “Robust Kalman filtering via Krein space estimation”,
IEE Proc.-Control Theory Appl., Vol. 151, No. 1, January 2004.
[11 Yvan Petillot, Ioseba Tena Ruiz, and David M. Lane, “Underwater Vehicle Obstacle Avoidance And
Path Planning Using a Multi-Beam Forward Looking Sonar,” IEEE journal of Oceanic Engineering,
vol.26, No.2, April 2001.
[12] Y. Bar-Shalom and T.E. Fortmann. Tracking and Data Association. Academic Press, 1988.
[13] E. Trucco, Y. Petillot, I. Tena Ruiz, C. Plakas, and D. M. Lane. Feature tracking in video and sonar
subsea sequences with applications. Computer Vision and Image Understanding, No.79, pages
92.122.16
[14] I. Tena Ruiz, D. M. Lane, and M. J. Chantler. A comparison of inter-frame feature measures for
robust object classi_cation in sector scan sonar image sequences. IEEE Journal of Oceanic
Engineering, 24, No.4:458.469, 1999.
[15] Cushieri J. and Negahdaripour S., “Use of forward scan sonar images for positioning and navigation
by an AUV”, in Proceedings of OCEANS’98, Nice, France, (IEEE/OES, Ed.), Vol. 2, pp:752–756,
September 1998.
[16] Bell J.M., Dura E., Reed S., Petillot Y.R., and Lane D.M., “Extraction and Classification of Objects
from Sidescan Sonar”, IEE Workshop on Nonlinear and Non-Gaussian Signal Processing, 8-9th July
2002.
[17] Clark D., Ruiz l.T., Petillot Y. and Bell J., “Multiple Target Tracking and Data Association in Sonar
Images”, The IEE Seminar on Target Tracking: Algorithms and Applications 2006 (Ref. No.
2006/11359), Birmingham, UK, pp: 147- 154, March 2006.
[18] Clark D.E. and Bell J., “Bayesian multiple target tracking in forward scan sonar images using the
PHD filter”, Radar, Sonar and Navigation, IEE Proceedings, Vol.152, Issue 5, pp:327 – 334, October
2005.

78
Authors
Smt Nagamani Modalavalasa received B.E. degree in Electronics and
Communication Engineering from Andhra University, Visakhapatnam, Andhra
Pradesh, India. She has completed her Master degree from Andhra University,
Visakhapatnam, India. At present she is a research scholar in Electronics &
Communication Engg. Department, JNTU Engg. College, Kakinada, Andhra
Pradesh, India. She has 15 years of teaching experience as Lecturer in the Department
of Electronics & Communication Engg, State Board of Technical Education &
Training, Andhra Pradesh, India. She has published around 15 research papers in
various International and National conferences.
Dr G Sasi Bhushana Rao received B E. degree in Electronics and Communication
Engineering from Andhra University College of Engineering, Visakhapatnam,
Andhra Pradesh, India, M.Tech. degree from JNTU, Hyderabad, India and Ph.D.
from Osmania University, Hyderabad, India. He possesses vast administration,
teaching and R&D experience at Airports Authority, Andhra University, India for
about 25 years. Currently he is working as Head Of Department in the Department of
Electronics & Communication Engg, Andhra University Engineering College,
Visakhapatnam, India. He has published more than 230 Technical and research
papers in different National / International conferences and Journals and authored two Text books. He has
guided 4 Ph.D. scholars and at present 14 scholars are working with him. His areas of Research include
Inertial Navigation System (INS), GPS/GNSS Signal processing, Ionosphere/Troposphere and Multipath
error modeling, and RADAR and SONAR navigation. Dr. Rao is a Fellow member of various professional
bodies like IEEE, IETE , IGU , and International GNSS society.
Dr. K. Satya Prasad received B Tech. degree in Electronics and Communication
Engineering from JNTU college of Engineering, Anantapur, Andhra Pradesh, India,
M.E. degree in Communication Systems from Guindy college of Engg. , Madras
University, Chennai, India and Ph. D from Indian Institute of Technology, Madras.
He has more than 31 years of experience in teaching and 23 years of R & D. He
started his teaching carrier as Teaching Assistant at Regional Engineering College,
Warangal in 1979. He joined JNT University, Hyderabad as Lecturer in 1980 and
served in different constituent college’s viz., Kakinada, Hyderabad and Anantapur and
at different capacities viz., Associate Professor, Professor, and Head of the
Department, Vice Principal and Principal. He has published more than 50 technical papers in different
National / International conferences and Journals and authored one Text book. He has guided 4 Ph.D.
scholars and at present 12 scholars are working with him. His areas of Research include Communications
Signal Processing, Image Processing, Speech Processing, Neural Networks & Ad-hoc wireless networks
etc. Dr. Prasad is a Fellow member of various professional bodies like IETE, IE, and ISTE.

AN EFFICIENT IMPLEMENTATION OF TRACKING USING KALMAN FILTER FOR UNDERWATER ROBOT APPLICATION

Recommended

More Related Content

What's hot (19)

Similar to AN EFFICIENT IMPLEMENTATION OF TRACKING USING KALMAN FILTER FOR UNDERWATER ROBOT APPLICATION (20)

More from IJCSEIT Journal (20)

Recently uploaded (20)

AN EFFICIENT IMPLEMENTATION OF TRACKING USING KALMAN FILTER FOR UNDERWATER ROBOT APPLICATION