Localization using filtered dgps

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 04 Issue: 04 | Apr-2015, Available @ http://www.ijret.org 471
LOCALIZATION USING FILTERED DGPS
Arun Kumar G C1
, Varalakshmi K R2
1
Asst Prof, Dept of ECE – SVIT (affiliated to VTU - Belgaum), Bangalore, Karnataka, India
2
Asst Prof, Dept of ECE – SVIT (affiliated to VTU - Belgaum), Bangalore, Karnataka, India
Abstract
In this paper, localization of the robot is achieved by considering two Global Positioning Systems (GPS) or DGPS. Differential
Global Positioning System (DGPS) is interfaced with MBED with the help of Zigbee protocol. For accurate localization of mobile
robot DGPS is preferred. Filters are used to remove the erroneous noise from the data obtained from GPS. Low pass IIR filter for
DGPS is realized. The project work discusses each of these approaches for localization in Outdoor environment. The above
algorithm is implemented on MBED Platform. Simulation results are extracted using Matlab.
Keywords—localization, Outdoor environment, Low pass IIR filter, DGPS, MBED
-------------------------------------------------------------------***-------------------------------------------------------------------
1. INTRODUCTION
GPS are used to estimate localization with respect to earth‟s
latitude, longitude and altitude axis. In earlier days
Researchers used to manually feed the localization data for
the robot. Later after the invention of GPS, Researchers
made use of GPS data in order to estimate localization [1]
[8]. The GPS all alone is of less accurate and is prone to
noise. Since there is random fluctuation in localization
update, later the researchers have came up with an idea of
DGPS, in which two GPS devices were used together. First
GPS sensor was placed on ground point (reference point)
and is interfaced with first MBED and the second GPS was
placed on robot and is interfaced with second MBED. For
communication between the two GPS, Zigbee (XBEE
Company) protocol is preferred [2].
Platform – Embedded Platform
Fig 1 MBED ARM Cortex-M3 LPC1768
MBED LPC1768 ARM Cortex-M3 is preferred as a
controller because of its high performance, low power
consumption. It supports I2
C, SPI, USB, CAN, Ethernet;
online compiling makes the programming task much easy.
GPS module is connected to ARM Micro Controller -
MBED [1]. Researchers have come up with many ideas for
estimating localization and many has been proposed in the
field of research in robotics. In earlier days for estimating
localization Ultrasonic positioning systems and infrared
network systems which are available at low cost, small
were used which can be easily interfaced with the
controller.
However, if the target is situated far, ultrasonic sensor fails
to detect the object; hence those kinds of systems were not
preferred. Due to large signal interference it is not preferred
nowadays. Later they have come up with an idea of Radio
Frequency Identification (RFID), which requires additional
equipment which was tied on robot; RFID requires high
cost to implement and hence it is also not preferred. For
Indoor navigation Global Positioning System (GPS) cannot
be preferred, as it requires Line of Sight (LoS) with the
satellites and has a slow update rate of one second
approximately. So, the researchers have come up with a
new idea of using Differential GPS which has an accuracy
level close to few tens of centimeter and are available and
used in commercial purpose and in vehicles also which
makes them attractive for localization of outdoor mobile
robots navigation [1] [2] [4].
Figure2 describes the project flow in which the project is
executed. GPS is connected to MBED; GPS device requires
power supply and is powered by MBED micro controller.
Tx and Rx pins of MBED are connected to Rx and Tx pins
of GPS for serial communication. As GPS data is prone to
noise, accuracy is in terms of few hundreds of meters.
Hence DGPS data is preferred which has a accuracy level
of few centimeters. On the application point of view robots
should navigate to destination accurately, filters are
preferred basically first or second order filter. For
simplicity first order is chosen.

_______________________________________________________________________________________
As two micro controllers are not wired, Zigbee protocol is
preferred for wireless communication between the two µC.
Fig 2 Flowchart of the design
Figure 2 describes the design flow of the paper. First GPS
devices are connected with MBEDs and the communication
between two MBED is done using Zigbee protocol (XBEE-
Company). The data obtained from GPS is erroneous to
noise, so appropriate filter algorithm is chosen according to
the design constraints.
2. FILTER ALGORITHM DESIGN
2.1 IIR Low Pass Filter
Infinite Impulse Response (IIR) is used to suppress or
eliminate the unwanted rapid fluctuation of GPS data. The
complexity of the system directly depends on the order of
the system. Gain „β‟ of the system decides the quantity of
noise to be removed from noise and „β‟ decides the
percentage of present data to be considered and (1- β)
determines the fraction of past data to be considered.
0<β<1
Output of the low pass Infinite Impulse Response (IIR)
filter equation for the first order system can be written as:
b[i] = (1-β)*b[i-1] + β * (c[i] - b[i-1]) (1)
where
c[i] = present input
b[i] = present output
b[i-1] = previous output
α = gain parameter
GPS data will provide information regarding latitude (lat),
longitude (long) and altitude (alt), N/S direction and also
E/W direction information. Localization of reference point
can be estimated by using set of eqns. of eqn. 2a, eqn. 2b &
eqn. 2c from GPS sensor.
Similarly Localization of mobile robot with respect to
reference point can be estimated by using set of eqns. of
eqn. 3a, eqn. 3b & eqn. 3c from GPS sensor.
Set of eqns. of 4 determines the gain quantity which is
dependent on both input and output. „β‟ on input and „1-β‟
on output.
From first GPS sensor extracting only latitude, longitude
and altitude information, these values are converted into
XYZ coordinates and is given by,
X1 = R*cos(lat1)*cos(long1) (2a)
Y1 = R*cos(lat1)*sin(long1) (2b)
Z1 = R*sin(alt1) (2c)
From second GPS extract latitude, longitude and altitude
information. Converting into coordinates:
X2 = R*cos(lat2)*cos(long2) (3a)
Y2 = R*cos(lat2)*sin(long2) (3b)
Z2 = R*sin(alt2) (3c)
First MBED sends GPS data (latitude, longitude and
altitude) into second MBED via Zigbee protocol
Since DGPS is used for calculating the localized
coordinate‟s difference of second GPS coordinates with the
first is taken. Localization of mobile robot considering
localization estimate information from set of eqn.2 and
eqn.3 with respect to ground point and is given by,
€X = X2-X1 (4a)
€Y = Y2-Y1 (4b)
€Z = Z2-Z1 (4c)
For serial communication between GPS devices and MBED
micro controller, BAUD rate was chosen to be 9600. If lock
is enabled, localized data is fetched from the satellite by
GPS sensors.
3. IMPLEMENTATION DESIGN
First GPS senses the north pole of earth (Heading
measurement). GPS sensor receives around 5 sets of data
stream ($GPGSV, $GPGGA, $GPGSA, $GPVTG
$GPRMC). For the application point of view $GPGGA is
chosen. As $GPGGA data stream provides information
regarding latitude, altitude, longitude information and it
also provides information regarding E/W direction and N/S
direction. [1] [3].
Fig 3 Block diagram representing communication between
two MBED's

_______________________________________________________________________________________
By which localization of robot can be estimated using
earth‟s coordinates in terms of X, Y, Z and the obtained
data is fed to low pass IIR design. Since data obtained from
As GPS sensor is prone to noise and accuracy is very much
poor, Differential GPS (DGPS) is chosen or preferred.
From figure.3, two GPS devices are connected to two
different micro controllers, so for communicating between
two microcontrollers Zigbee protocol (XBEE Company)
which supports wireless communication is chosen
(preferred).
4. SIMULATION RESULTS
Matlab tool is used to extract Simulation results. Mobile
robot is navigated in outdoor environment and localization
points are plotted with the help of Matlab tool and are
simulated to obtain the plots.
Fig 4 Filtering of latitude data obtained from GPS
Figure 4 describes the latitude variation with respect to
time. When the robot is navigating towards the destination,
due to rapid fluctuation in localization, in order to suppress
rapid fluctuation of the data accomplished with error signal,
filters are used.
Fig 5 Localization along x axis, β=0.5
Figure 5 describes the localized data along X axis (latitude
variation) and figure 6 describes longitude variation along
Y axis. Filter algorithm is done in order to smoothen the
waveform or to suppress rapid fluctuation in localized data.
Fig 6 Filtering of altitude data obtained from GPS
Fig 7 Localization along XYZ axis versus time (3D plot)
Figure.7 describes the localization estimate along X axis, Y
axis and Z axis with and without filter. Gain „β‟ was chosen
to be 0.5. (1-β) is the fraction of quantity depending on past
output, filtered response is plotted in figure 8.Gain (β) is
the fraction of quantity depending on input.

_______________________________________________________________________________________
Fig 8 Localization data with and without filter
5. CONCLUSION AND FUTURE WORK
The proposed filter algorithm for DGPS offers fast and
accurate localization in terms of earth co-ordinates and to
improve the localization and navigation of mobile robot in
outdoor environment or uneven surfaces accurately, multi-
sensor (camera, pressure sensor, IMU) sensor data can be
used. For Indoor navigation and in the absence of LoS,
IMU and Odometry data can be considered.
ACKNOWLEDGEMENTS
The authors express their sincere gratitude to Prof. M R
Holla - Director, Dr. Ramesh Babu H S - Principal, Prof. Y
Jayasimha – Dean Academics, Dr K Narayan – Dean R&D,
Dr. Chitra Kiran N – HOD, Dept of ECE, SVIT. The
authors would like to thank the faculties those who have
supported directly or indirectly in carrying out the research
work that was carried out in SVIT. The authors would also
wish to thank their family members for constant
encouragement and their support.
REFERENCES
[1] Arun Kumar G C, Chandrappa, Anil Kumar G C,
“Real time filter and fusion of multi-sensor for
localization of mobile robot”, IJECET, Volume 6,
Issue 1, January 2015, India, pp. 49 - 57
[2] Alejandro J. Weinstein, Kevin L. Moore, “Pose
Estimation of Ackerman Steering Vehicles for
Outdoors Autonomous Navigation” Golden, CO,
2010
[3] Eric North, Jacques Georgy, “Improved Inertial/
Odometry/ GPS positioning of wheeled robots even
in GPS denied Environments, 2012
[4] Armando AlvesNeto, Douglas Guimaraes Macharet,
“Adaptive complementary filtering algorithm for
mobile robot localization”, 2009
[5] Aeronautical Telecommunications, Annex 10 to the
Conversion on International Civil Aviation, 1985
[6] R. Hatch K P Schwarz and G Lachapelle,
“Kinematic systems in geodesy, surveying and
remote sensing”, pp.299 -308 1990
[7] M. S. Braasch "Improved Modeling of GPS
Selective Availability", Institute of Navigation 1993
National Technical Meeting, 1993
[8] Samadzadegan, FAbdi, ”Autonomous navigation of
Unmanned Aerial Vehicles based on multi sensor
data fusion”, 20th Iranian Conference on Electrical
Engineering (ICEE), pp. 868 – 873, 2012
[9] H. Wen, P. Y. Huang, J. Dyer, A. Archinal and J.
Fagan, "Countermeasures for GPS signal spoofing",
Proc. of the 18th
International Technical Meeting of
the Satellite Division of The Institute of Navigation,
pp. 1285-1290, 2005.
[10] P. Misra, P. Enge, “Global Positioning System,
Signal Measurements and Performance” 2ed,
Ganga-Jamura Press, Lincoln, 2006
BIOGRAPHIES
Arun Kumar G C, Working as an
Assistant Professor in SVIT, Bangalore.
He has finished his M.Tech (VLSI
Design and Embedded Systems) from
NMIT – Bangalore. He was also a
sponsored student from TEQIP during his
post graduation. He has worked in Centre
of Research in Robotics for 18 months
during his post graduation. He has completed his
graduation (B.E) from SVIT – Bangalore. He has won
second place for Mobile based car parking for SKYBUS in
International level competition held at Avadi, Chennai in
2012 during his graduation. His area of interest includes
Robotics, VLSI, Embedded System Design / Development.

_______________________________________________________________________________________
Varalakshmi K R, Working as an Assistant Professor in
SVIT, Bangalore. She has finished her M.Tech (Digital
Communication and Networking) from SJCIT,
Chikkabalapura. She has completed her graduation (B.E)
from NCET, Bangalore. She is having a teaching
experience of 5 years. Her area of interest includes
Embedded System design, Development, Networking. She
has now registered for PhD in VTU university.

Localization using filtered dgps

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Localization using filtered dgps (20)

More from eSAT Journals (20)

Recently uploaded (20)

Localization using filtered dgps