International Journal of Electrical & Electronics · PDF fileInternational Journal of...

International Journal of Electrical & Electronics Engineering Jan-Feb, 2014

Volume 1 Issue-1 IJEEE: e-ISSN: 1694-2310 | p-ISSN: 1694-2426

From Editor's Desk

Dear Reader,

I take the privilege to welcome all of you to this issue of

IJEEE. The intention of our journals is to create an

atmosphere that stimulates creativity, research and

development in the respective area of study. The development and growth of the mankind is the

consequence of brilliant researches done by scientists

and Engineers in every field. Our journals provides an

outlet for research findings and reviews in areas of

engineering, technology, science, management, law,

literature & philosophy found to be relevant for national

and international development. The aim and scope of

the journals is to provide an academic medium and an

important reference for the advancement and

dissemination of research results that support high-

level learning, teaching and research in the concerned

domains and to become a virtual knowledge grid for

researchers, scholars, postgraduate students, developers

and innovators in the respective fields.

Finally, I express my sincere gratitude to the Editorial

Board Members for their continuous support to make

the release of this special issue of the journals possible

and bring it to its present form. Undoubtedly we could

not issue an outstanding journal without the amazing

efforts of our reviewers. The motivation and expertise of

the professionals who serve as reviewers are necessary

to maintaining technical and editorial standards. So it is

with great pleasure and thankfulness that I offer my

thanks to each of the peer reviewers.

I hope you will enjoy reading this issue and we

welcome your feedback on any aspect of the Journal.

Naveen Kumar Editor-in-Chief,

Elixir Publications (Journals) [email protected]

IJEEE & IJCSIT is indexed by :

:

Disclaimer The responsibility of the contents and the opinions expressed in this journal is exclusively of the author(s) concerned. The publisher/editor of the journal is not responsible for errors in the contents or any consequences arising from the use of information contained in it. The opinions expressed in the research papers/articles in this journal do not necessarily represent the views of the publisher/editor.

www.ijeee-apm.com International Journal of Electrical & Electronics Engineering 1

IJEEE, Vol. 1, Issue 1 (Jan-Feb 2014) e-ISSN: 1694-2310 | p-ISSN: 1694-2426

DESIGN OF A RECTANGULARPATCH ANTENNA

Amandeep Bath1, Abhishek Thakur2, Jitender Sharma3, Prof. Basudeo Prasad4

1,2,3,4Electronics & Communication Department, IGCE, Punjab, [email protected], [email protected],[email protected]

Abstract- This paper presents the idea of recentdevelopments and advancements in the field of wirelesstechnology to realize high speed communications whichis performed in wideband technology .In this paper thewideband patch antenna is designed and fabricated. Apatch antenna is a narrowband, wide beam antennawhich is fabricated by etching the antenna elementpattern in metal trace bonded to an insulating dielectricsubstrate, such as a printed circuit board, in which acontinuous metal layer bonded to the opposite side ofthe substrate and it produces a ground plane. Thesimulation is done using ANSOFT HFSS simulationsoftware.

Index Terms- Ultra Wide Band (UWB); WirelessLocal Area Network (WLAN); Industrial, Scientific,and Medical Band (ISMB).

I. INTRODUCTION

In radio telecommunications, among the antennadesigns there are many different categories of microstrip antennas which are also known by the name printedantennas) the most common of which is the micro strippatch antenna or patch antenna. A patch antenna (alsoknown as a rectangular micro strip antenna) is a type ofradio antenna with a low profile, which can beconstructed on a flat surface. It consists of a flatrectangular metallic sheet or "patch" of metal, mountedover a larger metallic sheet called a ground plane. Theassembly is usually covered by a plastic radome, whichsaves the antenna structure from damage. Patchantennas are very simple to be fabricated and easy to bemodified and customized. They are the original type ofmicro strip antennas which were given by Howell in theyear 1972 in which the two metal sheets togetherproduce resonance and form a resonant piece of microstrip transmission line with a length which is around onehalf wavelength of the radio waves. The radiationprocess arises from discontinuities or irregularities ateach truncated edge of the micro strip transmission line.The radiation produced at the edges causes the antennato act slightly bigger electrically than its actual physicaldimensions, so in order for the antenna to be aresonanting piece of element, a length of micro striptransmission line slightly lesser than one half awavelength at the frequency is taken. A dielectricsubstrate is used for the construction of patch antenna,using the same. The easiest and most simple patchantenna uses a patch which is one half wavelength long,created at a precise distance above a larger groundplane, using an intermediary such as a spacer made of adielectric material between them. Electrically large

ground planes produce very rugged and stable patternsand lower environmental sensitivity but of courseincrease the size of the antenna. It isn’t uncommon forthe ground plane to be only slightly larger than theactive patch. When a ground plane is near to the size ofthe radiator it can have the phenomenon of coupling andproduce currents along the edges of the ground planewhich also radiate. The antenna pattern is created as thecombination of the two sets of radiating metallicelements. The current which flows is along the directionof the feed wire, so the magnetic vector potential andalso the electric field follow the current. A simple andeasiest patch antenna of this category radiates a linearlypolarized wave. The radiation can be considered asbeing produced by a number of the “radiating slots’’ attop and bottom, or simultaneously as a result of thecurrent flowing on the patch and the ground plane.Commonly made micro strip antenna shapes are square,rectangular, circular and elliptical, but any continuousshape is possible and can be created. Some patchantennas do not use a dielectric substrate and instead aremade by using a metal patch mounted above a groundplane using dielectric spacers; the resulting structure isless rugged but the bandwidth is much wider. Now assuch antennas have a very low profile, are mechanicallyrugged and can be shaped and designed to conform tothe curving skin of a vehicle, they are often mounted onthe exterior of aircraft and spacecraft, or areincorporated and operated into electronic devices suchmobile radio communication equipments.[1-6]

Antenna's features such as frequency, radiationpattern and polarization are reconfigured to achieve thedemands for agile radio applications. A lot of researchesfocus on frequency reconfiguration as futurecommunication systems such as cognitive radio needsan antenna that can do spectrum sensing andcommunication. In designing of reconfigurablefrequency antennas, recently a reconfigurable wide-bandto agile narrowband frequencies, using a printed logperiodic dipole array antenna, was developed. Awideband slotted antenna has been produced usingmultifunctional reconfigurable frequency characteristicsfor various applications such Wireless LAN, WIMAX,Ultra wideband and UMTS has been proposed in afrequency reconfigurable antenna, made up of twostructural elements ; one is an ultra-wide band (UWB)and other is a frequency reconfigurable triangularshaped antenna, is proposed for cognitive radioapplications

1) Ultra-wide band antennas have already been used inapplications such as satellite communication, remote

International Journal of Electrical & Electronics Engineering 2 www.ijeee-apm.com

sensing, ultra wide band radar technology and so on.Currently, the wireless area network (WLAN) in the 2.4-GHz (2.4-2.485 GHz) and 5-GHz (5.l5-5.875 GHz) bands isthe most renowned networks for accessing the internet andalso the antenna for an AP not only requires dual bandoperation but also needs to have an appropriate radiationprofile in both bands, namely equal gain, wide beam width,and high front-to-back ratio. Wireless communications isenjoying exponential growth in Industrial, Scientific, andMedical (ISM) band. The future generation wirelessnetworks require systems with broad band capabilities invarious environments to satisfy numerous applications assmart grid, personal communications, home, car, and officenetworking .On the other hand, many modern wirelesscommunication systems such as radar, navigation, satellite,and mobile applications use the circular polarized (CP)technology and radiation pattern. For the best UWBperformance, the transmitter and receiver (T/R) antennasshould have flat and high directive gain, narrow beam lowside and back lobes over the operational frequency band; toachieve the largest dynamic range, best focused illuminationarea, lowest T/R coupling, reduced ringing and uniformlyshaped impulse radiation.UWB has generally offers highdata rates at short distances with low power, primarily due tohigh resolution bandwidth.[7-11]

II. ANTENNA DESIGN AND CONFIGURATION

The geometry and configuration of the proposed antennais shown in the figure. Initially the design properties areselected by adjusting the local variables such as the substratethickness, height, material, transparency and position aswell. As shown in the figure the proposed antenna consistsof a substrate on which a cylindrical coax of Teflon isdeveloped. The cylindrical coax pin is made up of thematerial pec. Also the height and radius of the coax are –16.67mm and 0.283mm respectively. The feed pin is alsocylindrical with a radius of 0.083mm and the height of62mil. Before covering the design with a radiation air boxthe circular wave port on the substrate with a radius of0.283mm is made. Finally the design is covered with avacuum air box before the simulation and analysis.

Fig. 1: Rectangular patch antenna

Fig. 2: Air box over the antenna

The impedance bandwidth of a patch antenna is stronglyinfluenced by the spacing between the patch and theground plane. As the patch is moved closer to theground plane, less energy is radiated and more energy isstored in the patch capacitance and inductance: that is,the quality factor Q of the antenna increases.A patch printed onto a dielectric board is often moreconvenient to fabricate and is a bit smaller, but thevolume of the antenna is decreased, so the bandwidthdecreases because the Q increases, roughly in proportionto the dielectric constant of the substrate. Patch antennasutilized by industry often use ground planes which areonly modestly larger than the patch, which also alterstheir performance.[12-13]A. Rectangular patch

The most commonly designed micro strip antennais a rectangular patch. The rectangular patch antenna isaround a one half wavelength long strip of rectangularmicro strip transmission line. When air is taken as theantenna substrate, the length of the rectangular microstrip antenna is approximately one half of a free spacewavelength. Antenna is loaded with a dielectricsubstrate. The length of the antenna reduces as therelative dielectric constant of the substrate elementincreases. The resonant length of the antenna is slightlylesser because of the increased electric "fringing fields"which increase the electrical length of the antennaslightly. An old model of the micro strip antenna is asection of micro strip transmission line with equivalentloads on either end to represent the radiation loss.B. Planar inverted F antenna

Another category of patch antenna is the PlanarInverted-F Antenna (PIFA) common in cellular phoneswith built-in antennas. (The Planar Inverted-F antenna(PIFA) is highly used in the mobile phone market. Theantenna is resonant at a quarter-wavelength (thusdecreasing the required space needed on the phone), andalso typically has good SAR properties. This antennaresembles an inverted F, which explains the PIFA name.The Planar Inverted-F Antenna is renowned because ithas a low profile and an Omni directional pattern. Theseantennas are produced from a quarter wave half patchantenna. The shorting plane of the half-patch isdecreased in length which decreases the frequency ofresonance. Often PIFA antennas have multiple branchesfor resonating at the multiple cellular bands. On somephones, grounded parasitic elements are applied toimproves the radiation bandwidth characteristics. [14-16]C. Advantages


Micro strip antennas are comparatively inexpensiveto manufacture and fabricate because of the easy 2dimensional physical construction and geometry. Theyare usually employed at UHF and other higherfrequencies because the size of the antenna is directlyrelated to the wavelength at the frequency of resonance.A single patch antenna gives a maximum directive gainof around 6bB to 9dB. It is relatively easy to print anarray of patches on a single (large) substrate andlithographic techniques are used for this. Patch arrayscan give much higher gains than a single patch at littleadditional expense; matching and adjustment of phasecan be performed with printed micro strip feedstructures, again in the same operations that producesthe radiating patches. The capability to create high gainarrays in a low-profile antenna is one reason that patcharrays are commonly used on airplanes and in othermilitary applications. Such an array of patch antennas isan easy way to design a phased array of antennas withdynamic beam forming capability. An advantageinherent to patch antennas is the skill to havepolarization diversity. Patch antennas can easily befabricated to have vertical, horizontal, right handcircular (RHCP) or left hand circular (LHCP)polarizations or different kinds of polarizations, usingmultiple feed points, or a single feed point withasymmetric patch structures. This unique feature enablespatch antennas to be used in many types ofcommunications links that may have variedrequirements.

III. PATCH ANTENNA DESIGN CONSIDERATION

Wideband antenna are designed and fabricated forsmart grid applications with a frequency bandwidth of40% and gain of 3db to 4db . The antenna design andsimulation was carried out using ANSYS’ HFSS that isthe high frequency structure simulator software which isthe industry standard simulation tool for the simulationof 3D full wave electromagnetic field. The mostcommonly employed micro strip antenna is arectangular patch. The rectangular patch antenna isaround a one half wavelength long strip of rectangularmicro strip transmission line. When air is kept as theantenna substrate, the length of the rectangular microstrip antenna is approximately one half of a free spacewavelength. The length of the antenna reduces as theantenna is loaded with a dielectric as its substrate aswell as the relative dielectric constant of the substratehighly increases. The resonant length of the antenna isslightly lesser because of the incerased electric "fringingfields" which improve the electrical length of theantenna a little. An old model of the micro strip antennais a strip of micro strip transmission line with equalloads on either end to represent the radiation losses.

Fig. 3: Patch antenna design

It is possible to fabricate patch antennas that radiatewaves which are circularly polarized. One approach is toexcite a single square patch using two feeds, with onefeed with a phase difference of 90° with respect to theother. This drives each transverse mode and with equalamplitudes and the required 90 degrees out of phase.Each mode radiates separately and combines to producecircular polarization. This feed condition is oftenavailable using a 90 degree hybrid coupler. When theantenna is fed in this manner, the vertical current flow ismaximized and is high as the horizontal current flowbecomes zero, so the radiated electric field will bevertical; one quarter cycle later, the situation will havereversed and becomes opposite and the field will behorizontal. The radiated field rotates in time, producinga circularly polarized wave. An alternative is to use asingle feed but introducing some sort of asymmetric slotor other feature on the patch, causing the currentdistribution to be completely displaced. A square patchwhich has been perturbed slowly to produce arectangular micro strip antenna can be driven along adiagonal and create circular polarization. The aspectratio of this rectangle is chosen so each orthogonal modeis both non resonant. At the driving point of the antennaone mode is +45 degrees and the other -45 degrees andit is required to produce the 90 degree phase shift forcircular polarization.[16-20]

IV. RESULTS AND DISCUSSIONNow In this section the rectangular patch antenna is

designed and the numerical and experimental resultsregarding the radiation characteristics are presented anddiscussed. The simulated results are obtained by usingthe Ansoft simulation software high frequency structuresimulator. The measured and simulated characteristicsof the antenna are shown from the far field report of therectangular plot, the 3D polar plot and radiationcharacteristics.

Fig. 4: Return loss of antenna

A diagrammatic radiation pattern for a linearly-polarized 900MHz patch antenna is presented below.The figure shows a cross-section of a horizontal plane;the pattern in the vertical plane resembles though it isnot exactly identical. The scale is logarithmic, so thepower radiated at 180° (90° to the left of the beamcenter) is about 15 dB less than the power in the centerof the beam. The beam width is around 65° and the gainis approximately 9dB. An infinitely-big ground planeprevents any radiation which comes to the back of theantenna (angles from 180 to 360°), but the actualantenna has a fairly short ground plane, and the power


which comes in the backwards direction is only around20 dB down from that in comparison to the main beam.

Fig. 5: Radiation pattern of antenna

Fig. 6: Input impedanceof antenna

Unlike other antennas mentioned in literature todate, the proposed antenna presents a good Omnidirectional radiation pattern even at the very highfrequencies. The designed antenna has a very small sizeand even the return loss is low and radiation patterncharacteristics are obtained in the frequency band whichis used. The simulated and experimental results givesthe idea how that the proposed antenna could be a goodcandidate for UWB applications.

Fig. 7:ff_2D_Gain totalof antenna

Fig. 8: Return lossof antenna

As the length of the patch, half a wavelength, isabout the similar as the length of a resonating dipole, weget about 2 dB of gain out of the directivity relative tothe vertical axis of the patch antenna. If the patch iscompletely square, the pattern in the horizontal planewill be directional, somewhat as if the patch were a pairof dipoles which were separated by a half-wave; thiscounts for about another2-3 dB. Finally, the addition ofthe ground plane removes most or all radiation behindthe antenna, decreasing the power averaged over alldirections by a factor of around 2 (and therebyincreasing the gain by a factor of 3 dB). Summing thisall up, we get about 7-9 dB for a square element patch,in good agreement with more recognized approaches

Fig. 9: 3D polar plotof antenna

V. CONCLUSSION AND FUTURE WORKWith the rapid development of wireless

technology in recent years, various wireless systemssuch as GSM, WCDMA/UMTS, Bluetooth, WLANs,and GPS have been highly integrated into the mobileequipments, and in order to fulfill the RF systemrequirements using the different frequency band,antenna technology is required to widebandcharacteristics .On the other hand, many modernwireless communication systems such as radar,navigation, satellite, and mobile applications use thecircular polarized (CP) radiation pattern. The attractivefeatures of the CP antenna are existed as follows.Firstly, since the CP antennas transmit and receive in allplanes all around, it is strong for the reflection andabsorption of the radio signal. In the multi-path fadingchannel environment, the CP antenna overcomes out ofphase problem which can create dead-spots, decreasedthroughput, reduced overall system performance.Further advancements could be done by using antennasubstrates with higher dielectric constants in order tominimize the size a broad band wide beam circular


polarization micro strip antenna. The configuration ofthe antenna is the simplest and easiest to be fabricatedas compared with conventional micro strip antenna, theradiation beam is broadened a lot. Further research oncircularly polarized wideband micro strip antenna isrequired as it gives the best performance and overallimprovement of antenna parameters.

REFERENCES[1] Gaboardi P., Rosa L., Cucinotta A., and Selleri S., “Patch Array

Antenna for UWB Radar Applications”, in 3rdEuropean RadarConference, 2006, p.281-284.

[2] Letestu and Ala Sharaiha, “Size reduced multi-band printedquadrifilar helical antenna,” IEEE Trans. Antennas Propag.,vol. 59, pp. 3138-3143, 2011.

[3] A. Siligaris et al., “A 65-nm CMOS fully integrated transceivermodule for 60-GHz Wireless HD applications,” IEEE Journalof Solid-State Circuits, vol. 46, no. 12, pp. 3005-3017, Dec.2011.

[4] S. Manafi, S. Nikmehr, and M. Bemani, "Planar reconfigurablemultifunctional antenna for WLAN/ wimax/ UWB/ pcsdcs/UMTS applications," Progress In Electromagnetic Research C,Vol. 26, 123- 137, 2012.

[5] F. Ghanem, P. S. Hall and J. R. Kelly, “Two port frequencyreconfigurable antenna for cognitive radios”, ElectronicsLetters,vol.45, 2009,pp.534-536.

[6] C. R. Medeiros, E. B. Lima, 1. R. Costa, and C. A. Fernandes,"Wideband slot antenna for WLAN access point," IEEEAntenna Wireless Propagate. Lett., vol. 9,pp. 79-82, 2010.

[7] E. Ebrahimi, J. R. Kelly and P. S. Hall, “A reconfigurableNarrowband antenna integrated with wideband monopole forcognitive radio applications”, IEEE Antennas and PropagationSociety International Symposium (APSURSI), 2009.

[8] J. W. Baik, S. Pyo, T.H. Lee, and Y.S. Kim, “Switchableprinted Yagi Uda antenna with pattern reconfiguration”, ETRIJournal, vol.31 2009,pp.318-320.

[9] M. Sanad, "A Small Size Micro strip Antenna Circuit", IEEEInternational Conference on Antenna and Propagation, vol. 1,pp. 465-471, April1995.

[10] P. Suraj and V. R. Gupta, “Analysis of a Rectangular MonopolePatch Antenna” International Journal of Recent Trends inEngineering, Vol. 2, No. 5, pp. 106-109, November 2009.

[11] M. N. Srifi, M. Meloui and M. Essaaidi, “Rectangular SlottedPatch Antenna for 5-6GHz Applications”, International Journalof Microwave and Optical Technology, Vol.5 No. 2, pp., 52-57March 2010.

[12] Ansoft Corporations, HFSS V.12- Software based on the finiteelement method.

[13] G. Augustin, S. V. Shynu, C. K. Aanandan, and K. Vasudevan,"Compact dual band antenna for wireless access point, "Electron. Lett., vol. 42, no. 9, pp. 502-503, Apr. 2006.

[14] S. W. Su, "Concurrent dual-band six-loop-antenna system withwide 3-dB beam width radiation for MTMO access point, "Microwave Opt. Techn. Lett., vol. 52, no. 6, pp. 1253-1258,Jun. 2010.

[15] S. W. Su, "High-gain dual-loop antenna for MTMO accesspoint in the 2.4/5.2/5.8 GHz bands, " IEEE Trans. AntennaPropag., vol. 58, no. 7, pp. 2412-2419, Jul. 2010.

[16] K.B. Hsieh, M.H. Chen, and K.L. Wong, "Single feed DualBand Circularly Polarized Micro strip antenna," Electron. Lett.VOL. 34, pp.1170-1171,Jun.1998

[17] G. P. Jin, D.L. Zhang and R.L. Li, “Optically controlledReconfigurable antenna for cognitive radios applications”,Electronics Letters, vol.47, pp.948-950, 2011.

[18] Wang Yazhou, Su Donglin, Xiao Yongxuan, “Broadbandcircularly polarized square micro strip antenna”, Antennas,Propagation and EM Theory, 2006, ISAPE 2006 7thInternational Symposium on, pp.1–4,2006.

[19] LTCC hybrid coupler, “RCP1850S03N”, RN2 Technologies co.

Zhen-Yu Zhang, Yong-Xin Guo, Ling Chuen Ong, ChiaM.Y.W., “A New wide-band planar balun on a single layerPCB”, Microwave and Wireless Components Letters, IEEE,Vol. 15, No. 6, pp. 416–18, Jun. 2005.

AUTHORSFirst Author – AmandeepBatth:M. Tech. in Electronics andCommunication Engineeringfrom Punjab TechnicalUniversity, MBA in HumanResource Management fromPunjab Technical University ,Bachelor in Technology (B-Tech.)from Punjab Technical

University . Six years of work experience in teaching.Area of interest: Antenna Design and WirelessCommunication. International Publication: 1, NationalConferences and Publication: 4. Working with IndoGlobal College of Engineering Abhipur, Mohali, P.B.since 2008.Email address: [email protected]

Second Author– AbhishekThakur: M. Tech. in Electronicsand CommunicationEngineering from PunjabTechnical University, MBA inInformation Technology fromSymbiosis Pune, M.H.Bachelor in Electronics (B.E.)from Shivaji UniversityKolhapur, M.H. Five years of

work experience in teaching and one year of workexperience in industry. Area of interest: Digital Imageand Speech Processing, Antenna Design and WirelessCommunication. International Publication: 7, NationalConferences and Publication: 6, Book Published: 5(Microprocessor and Assembly Language Programming,Microprocessor and Microcontroller, DigitalCommunication and Wireless Communication).Working with Indo Global College of EngineeringAbhipur, Mohali, P.B. since 2011.Email address: [email protected] Author – Jitender Sharma: M. Tech. inElectronics and Communication Engineering fromMullana University, Ambala, Bachelor in Technology(B-Tech.)from Punjab Technical University . Five yearsof work experience in teaching. Area of interest:,Antenna Design and Wireless Communication.International Publication: 1 National Conferences andPublication:6 and Wireless Communication). Workingwith Indo Global college since 2008E-mail address:[email protected]



DESIGN OF IMAGE COMPRESSIONALGORITHM USING MATLABAbhishek Thakur1, Rajesh Kumar2, Amandeep Bath3, Jitender Sharma4

1,2,3,4Electronics & Communication Department, IGCE, Punjab, [email protected], [email protected],

[email protected], [email protected]

Abstract- This paper gives the idea of recentdevelopments in the field of image security andimprovements in image security. Images are used inmany applications and to provide image security usingimage encryption and authentication. Image encryptiontechniques scramble the pixels of the image anddecrease the correlation among the pixels, such that theencrypted image cannot be accessed by unauthorizeduser. This study proposes a method for encrypting thesender’s messages using new algorithm called chaoticencryption method. This key will be used for encryptingand decrypting the messages which are transmittedbetween two sides. Chaotic encryption technique is thenew way of cryptography. Emphasizing the imagesecurity, this paper is to design the enhanced securealgorithm which uses chaotic encryption method toensure improved security and reliability.

Index Terms- Cryptography, Cryptography key,RGB image, LSB, image key.

I. INTRODUCTION

In telegraphy Morse code, is a data compressiontechnique using shorter codewords for letters invented in1838. In late 1940s information theory is invented. In1949 Claude Shannon and Robert Fano develop a wayto assign code words based on probabilities of blocks.Later an optimal method for doing this was developedby David Huffman in 1951. Early implementations weredone in hardware, with specific choices of codewordsbeing made as compromises between error correctionand compression. In the mid-1970s, Huffman encodingfor dynamically updating codewords based on the actualdata encountered is invented. Pointer based encoding isinvented in 1977 by Abraham Lempel and Jacob Ziv.LZW algorithm is invented in the mid-1980s, whichbecame the method for most general purposecompression systems. It was used in programs, as wellas in hardware devices such as modems [1]. In the early1990s, lossy compression method is widely used. Todayimage compression standards include: FAX CCITT 3(run-length encoding, with codewords determined byHuffman coding from a definite distribution of runlengths); BMP (run-length encoding, etc.); JPEG (lossydiscrete cosine transform, then Huffman or arithmeticcoding); GIF (LZW); TIFF (FAX, JPEG, GIF, etc.)[2].Typical compression ratios currently achieved for textare around 3:1, and for photographic images around 2:1lossless, and for line diagrams and text images around3:1, and 20:1 lossy. Images on internet grow very fast

and hence each server needs to store high volume ofdata about the images and images are one of the mostimportant data about information. As a result servershave a high volume of images with them and require ahuge hard disk space and transmission bandwidth tostore these images. Most of the time transmissionbandwidth is not sufficient for storing all the image data.Image compression is the process of encodinginformation using fewer bits. Compression is usefulbecause it helps to reduce the consumption of expensiveresources, such as transmission bandwidth or hard diskspace. Compression scheme for image may requireexpensive hardware for the image to be decompressedfast enough to be viewed as its being decompressed [3].Compression schemes therefore involves trade-offsamong various factors, including the degree ofcompression, the amount of distortion introduced, andthe computational resources required to compress anduncompress the data. Image compression is theapplication of Data compression. Image compressionreduce the size in bytes of a graphics file withoutdegrading the quality of the image. The purpose ofreduction in file size allows more images to be stored ina given amount of disk or memory space. It also reducesthe time required for images to be sent over the Internetor downloaded from Web pages. There are severaldifferent ways in which image files can be compressed.For Internet use, the two most common compressedgraphic image formats are the JPEG format and the GIFformat. The JPEG method is more often used forphotographs, while the GIF method is commonly usedfor line art and other images in which geometric shapesare relatively simple [4-6].

Other techniques for image compression includethe use of fractals and wavelets. These methods have notgained widespread acceptance for use on the Internet asof this writing. However, both methods offer promisebecause they offer higher compression ratios than theJPEG or GIF methods for some types of images.Another new method that may in time replace the GIFformat is the PNG format [7]. With a compression ratioof 32:1, the space, bandwidth and transmission timerequirements can be reduced by the factor of 32, withacceptable quality. Image compression and codingtechniques explore three types of redundancies:D. Coding redundancy:

Coding redundancy is present when less than optimalcode words are used. This type of coding is alwaysreversible and usually implemented using look up tables(LUTs) [8]. Examples of image coding schemes that


explore coding redundancy are the Huffman codes andthe arithmetic coding technique.E. Inter pixel redundancy:

This type of redundancy sometimes called spatialredundancy, inter frame redundancy, or geometricredundancy. This redundancy can be explored in severalways, one of which is by predicting a pixel value basedon the values of its neighboring pixels. In order to do so,the original 2-D array of pixels is usually mapped into adifferent format, e.g., an array of differences betweenadjacent pixels. If the original image pixels can bereconstructed from the transformed data set the mappingis said to be reversible. Examples of compressiontechniques that explore the interpixel redundancyinclude: Constant Area Coding (CAC), (1-D or 2-D)Run Length Encoding (RLE) techniques, and manypredictive coding algorithms such as Differential PulseCode Modulation (DPCM) [9].F. Psycho visual redundancy:

Many experiments on the psychophysical aspects ofhuman vision have proven that the human eye does notrespond with equal sensitivity to all incoming visualinformation; some pieces of information are moreimportant than others. The knowledge of whichparticular types of information are more or less relevantto the final human user have led to image and videocompression techniques that aim at eliminating orreducing any amount of data that is psycho visuallyredundant. The end result of applying these techniquesis a compressed image file, whose size and quality aresmaller than the original information, but whoseresulting quality is still acceptable for the application athand. The loss of quality that ensues as a byproduct ofsuch techniques is frequently called quantization, as toindicate that a wider range of input values is normallymapped into a narrower range of output values thoroughan irreversible process [10]. In order to establish thenature and extent of information loss, different fidelitycriteria (some objective such as root mean square(RMS) error, some subjective, such as pair wisecomparison of two images encoded with differentquality settings) can be used. Most of the image codingalgorithms in use today exploit this type of redundancy,such as the Discrete Cosine Transform (DCT)-basedalgorithm at the heart of the JPEG encoding standard.

II. TECHNIQUES USED FOR IMAGE COMPRESSION

A. Image Compression Model:Image compression system is composed of two

distinct functional components: an encoder and adecoder. The encoder performs the complementaryoperation of compression, and the decoder performs thecomplementary operation of decompression. Bothcompression and decompression operations can beperformed in software, as in the case in web browsersand many commercial image editing programs, or in acombination of hardware and firmware, as incommercial DVD players [11]. A codec is a device orprogram that is capable of both encoding and decoding.

Input image f(x, y) is fed into the encoder, whichcreates a compressed representation of the input. Thisrepresentation is stored for latter use, or transmitted forstorage and use at a remote location. When the

compressed representation is presentation is presented toits complementary decoder, a reconstructed outputimage f^(x, y) is generated. In general, f^(x,y) may ormay not be a replica of f(x,y). If it is, the compressionsystem is called error free, lossless or informationpreserving. If not, the reconstructed output image isdistorted and the compression system is referred to aslossy.

B. The encoding or compression process:The encoder is designed to remove the

redundancies. In the first stage of the encoding process,a mapper transforms f(x, y) into a format designed toreduce spatial and temporal redundancy. This operationgenerally is reversible and may or may not reducedirectly the amount of data required to represent theimage. The run length coding is an example of themapping that normally yields compression in the firststep of encoding process. The quantizer reduces theaccuracy of the mapper’s output in accordance with per-established fidelity criterion. The goal is to keepirrelevant information out of the compressedrepresentation. This operation is irreversible. It must beomitted when error-free compression is desired. In thethird and final stage of the encoding process, the symbolcoder generates a fixed or variable length code torepresent the quantizer output and maps the output inaccordance with the code [12]. In many cases, a variable–length code is used. The shortest code words areassigned to the most frequently occurring quantizeroutput values- thus minimizing coding redundancy. Thisoperation is irreversible. Upon its completion, the inputimage has been processed for the removal of each of thethree redundancies (Coding, Interpixel , Psycho visual).

C. The Decoding or decompression process:The decoder contains only two components: a

symbol decoder and an inverse mapper. They perform,in reverse order, the inverse operation of encoder’ssymbol encoder and mapper. Because quantizationresults in irreversible information loss, an inversequantizer block is not included in the general decodermodel [13].

D. Image Compression Algorithms:Image compression can be lossy or lossless.

Lossless compression is sometimes preferred forartificial images such as technical drawings, icons orcomics. This is because lossy compression methods,especially when used at low bit rates, introducecompression artifacts. Lossless compression methodsmay also be preferred for high value content, such asmedical imagery or image scans made for archivalpurposes. Lossy methods are especially suitable fornatural images such as photos in applications whereminor (sometimes imperceptible) loss of fidelity isacceptable to achieve a substantial reduction in bit rate[14-15].

E. Various Lossy Compression Methods are:2) Cartesian Perceptual Compression: Also known

as CPC3) DjVu


4) Fractal compression5) HAM, hardware compression of color

information used in Amiga computers6) ICER, used by the Mars Rovers: related to JPEG

2000 in its use of wavelets7) JPEG8) JPEG 2000, JPEG's successor format that uses

wavelets.9) JBIG210) PGF, Progressive Graphics File (lossless or lossy

compression) Wavelet compression.

F. Various Loss-Less Compression Method are:1) Run-length encoding – used as default method in

PCX and as one of possible in BMP, TGA, TIFF2) Entropy coding3) Adaptive dictionary algorithms such as LZW –

used in GIF and TIFF4) Deflation – used in PNG, MNG and TIFF.

G. The steps involved in compressing anddecompressing of image are:

1) Specifying the Rate (bits available) andDistortion (tolerable error) parameters for thetarget image.

2) Dividing the image data into various classes,based on their importance.

3) Dividing the available bit budget among theseclasses, such that the distortion is a minimum.

4) Quantize each class separately using the bitallocation information derived in step 3.

5) Encode each class separately using an entropycoder and write to the file.

6) Reconstructing the image from the compresseddata is usually a faster process than compression.The steps involved are step 7 to step 9.

7) Read in the quantized data from the file, using anentropy decoder. (Reverse of step 5).

8) Dequantized the data. (Reverse of step 4).9) Rebuild the image. (Reverse of step 2).

H. Calculation of image compression:There are various types of terms that are used in

calculation of image compression. Some are listedbelow:

a. Cartesian Perceptual Compression: Also known asCPC

b. Compression Ratio = UncompressedSize/Compressed Size

c. Space Saving = [1- compressed Size/UncompressedSize]

d. Data Rate Savings = [1- compressed datarate/Uncompressed data rate]

The PSNR is most commonly used as a measure ofquality of reconstruction in image compression etc. It ismost easily defined via the mean squared error (MSE)which for two m×n monochrome images I and K whereone of the images is considered a noisy approximationof the other is defined as:

1

0

1

0

),(),(1 m

i

n

j

jiKjiIMN

MSE 2

The PSNR is defined as:

MSE

MAX

MSE

MAXPSNR 1

10

21

10 log20log10

Here, MAXi is the maximum possible pixel value ofthe image. When the pixels are represented using 8 bitsper sample, this is 255. More generally, when samplesare represented using linear PCM with B bits persample, MAXI is 2B-1.Signal-to-noise ratio is a term for the power ratiobetween a signal (meaningful information) and thebackground noise:

2

Sinal

Signal

Noise

Signal

A

A

P

PSNR

If the signal and the noise are measured across thesame impedance then the SNR can be obtained bycalculating 20 times the base-10 logarithm of theamplitude ratio:

Noise

Signal

Noise

Signal

A

A

P

PdbSNR 1010 log20log10)(

The MSE of an estimator with respect to theestimated parameter θ is defined as:

MSE ()=E((-)2)The MSE can be written as the sum of the variance

and the squared bias of the estimator:MSE ()=Var()+(Bias(,))2

In a statistical model where the estimand isunknown, the MSE is a random variable whose valuemust be estimated. This is usually done by the samplemean, with θj being realizations of the estimator ofsize n.

2

1

'' )(1

)(

n

jjn

MSE

III. COMPRESSION TECHNOLOGIES

A. All Block Truncation Coding:Block Truncation Coding (BTC) is a well known

compression scheme proposed in 1979 for the greyscale.It was also called the moment preserving blocktruncation because it preserves the first and secondmoments of each image block. It is a lossy imagecompression technique. It is a simple technique whichinvolves less computational complexity. BTC is a recenttechnique used for compression of monochrome imagedata. It is one-bit adaptive moment-preserving quantizerthat preserves certain statistical moments of smallblocks of the input image in the quantized output. Theoriginal algorithm of BTC preserves the standard meanand the standard deviation. The statistical overheadsMean and the Standard deviation are to be coded as partof the block. The truncated block of the BTC is the one-


bit output of the quantizer for every pixel in the block[1].B. Wavelet Transform:

A wavelet image compression system can be createdby selecting a type of wavelet function, quantizer, andstatistical coder. The first step in wavelet compressionare performing a discrete wavelet Transformation(DWT), after that quantization of the wavelet spaceimage sub bands, and then encoding these sub bands.Wavelet images are not compressed images; rather it isquantization and encoding stages that do the imagecompression. Image decompression, or reconstruction,is achieved by carrying out the above steps in reverseand inverse order [4].C. EZW Algorithm:

In embedded coding, the coded bits are ordered inaccordance with their importance and all lower ratecodes are provided at the beginning of the bit stream.Using an embedded code, the encoder can terminate theencoding process at any stage, so as to exactly satisfythe target bit-rate specified by the channel. To achievethis, the encoder can maintain a bit count and truncatethe bit-stream, whenever the target bit rate is achieved.Although the embedded coding used in EZW is moregeneral and sophisticated than the simple bit-planecoding, in spirit, it can be compared with the latter,where the encoding commences with the mostsignificant bit plane and progressively continues withthe next most significant bit-plane and so on. If targetbit-rate is achieved before the less significant bit planesare added to the bit-stream, there will be reconstructionerror at the receiver, but the “significance ordering” ofthe embedded bit stream helps in reducing thereconstruction error at the given target bit rate [2].D. Fractal Image Compression and Decompression:

Fractal compression is a lossy image compressionmethod using fractals to achieve high levels ofcompression. The method is best suited for photographsof natural scenes (trees, mountains, ferns, clouds).The fractal compression technique relies on the fact thatin certain images, parts of the image resemble otherparts of the same image. Fractal algorithms convertthese parts, or more precisely, geometric shapes intomathematical data called "fractal codes" which are usedto recreate the encoded image. Fractal compressiondiffers from pixel-based compression schemes such asJPEG, GIF and MPEG since no pixels are saved. Oncean image has been converted into fractal code itsrelationship to a specific resolution has been lost. Theimage can be recreated to fill any screen size without theintroduction of image artifacts that occurs in pixel basedcompression schemes.

IV. RESULTSMatlab approach has been implemented for compressingimages. Matlab is capable of handling simultaneousmultiple matrix of images. In this firstly given RGBimage is divided into three matrices which arecompressed individually and then these are mergedtogether to construct final compressed image. Imageprocessing tool box of matlab is used to perform allcompression task.

Matlab based environment ensures high scalability,availability and reliability through redundancymechanisms. Hence matlab computing proves to be anappropriate platform for compressing images by varioustechniques. Matlab provides High-level language fortechnical computing and Development environment formanaging code, files and data. It is an Interactive toolfor iterative exploration, design and problem solving. Itprovides Mathematical functions for linear algebra,statistics, Fourier analysis, filtering, optimization,numerical integration and Tools for building customgraphical user interfaces.

In this study compressing of images is taken bysplitting image matrix into R,G,B matrix which arecompressed individual and then these are merged forconstruction of compressed image. It is an efficientapproach for performing compression for large sizeimages of different format. Time and space is reduced togreat extent with the help of Matlab computing.

An RGB image, sometimes referred to as a truecolor image, is stored as an m-by-n-by-3 data array thatdefines red, green and blue color components for eachindividual pixel. RGB images do not use a palette.

The colour of each pixel is determined by thecombination of the red, green, and blue intensities storedin each colour plane at the pixel's location. Graphics fileformats store RGB images as 24-bit images, where thered, green, and blue components are 8 bits each. Thisyields a potential of 16 million colours.

A. Block Truncation Coding:In table 1 we take the block size of 8*8, 64*64 andperform block truncation coding. These results areshown in figure 1.Table 1: Block Truncation Coding

Block Size 8*8 64*64MSE 3.5077e+003 3.2613e+003

PSNR 12.6805 12.9969

a)Original Image b) 8*8 Block Truncation c) 64*64 BlockTruncation

Fig. 1: Block Truncation Coding

B. Wavelet Compression:In table 2 Wavelet Compression encode time, MSEand PSNR is shown with Decomposition Level 3


and 7. Figure 2 shows results for WaveletCompression.

Table 2: Wavelet Compression

Decom.Level

4 7

Encodetime

35.188 Sec. 35.359 Sec.

MSE 3.5609e+003 3.6379e+003PSNR 12.6152 12.5224

a) Original b) Level 4 c) Level 7Fig. 2: Wavelet Compression

C. Embedded Zerotree:In table 3 Embedded Zerotree encode time, decodetime, MSE and PSNR is shown with decompositionlevel 4 and 6. Figure 3 shows the results forEmbedded Zerotree.

Table 3:Embedded Zerotree

Decom.Level

4 6

Encodetime

237.7180Sec.

243.8910Sec.

Decodetime

119.9070Sec.

123.4360Sec.

MSE 3.4228e+003 3.4228e+003PSNR 12.7870 12.7870

a) Original b) Level 4 c) Level 6Fig. 3: Embedded Zerotree

D. Fractal Image Compression:In table 4 Fractal Image Compression encode time,decode time, MSE and PSNR is shown with increasein block size of 10 and 25. Figure 4 shows the resultsfor Fractal Image Compression.


Decom.Level

4 25

Encodetime

300.0610Sec.

302.310 Sec.

Decodetime

41.0320 Sec. 40.703 Sec.

MSE 3.5205e+003 2.9751e+003PSNR 12.6648 13.3957

a) Original b) Block Size 2 c) Block Size25

Fig. 4: Embedded Zerotree

Fig. 5 Flowchart for implementation of compressing image in matlabEnvironment




Decom.Level

4 7

Encodetime

35.188 Sec. 35.359 Sec.

MSE 3.5609e+003 3.6379e+003PSNR 12.6152 12.5224




Decom.Level

4 6

Encodetime

237.7180Sec.

243.8910Sec.

Decodetime

119.9070Sec.

123.4360Sec.

MSE 3.4228e+003 3.4228e+003PSNR 12.7870 12.7870




Decom.Level

4 25

Encodetime

300.0610Sec.

302.310 Sec.

Decodetime

41.0320 Sec. 40.703 Sec.

MSE 3.5205e+003 2.9751e+003PSNR 12.6648 13.3957







Decom.Level

4 7

Encodetime

35.188 Sec. 35.359 Sec.

MSE 3.5609e+003 3.6379e+003PSNR 12.6152 12.5224




Decom.Level

4 6

Encodetime

237.7180Sec.

243.8910Sec.

Decodetime

119.9070Sec.

123.4360Sec.

MSE 3.4228e+003 3.4228e+003PSNR 12.7870 12.7870




Decom.Level

4 25

Encodetime

300.0610Sec.

302.310 Sec.

Decodetime

41.0320 Sec. 40.703 Sec.

MSE 3.5205e+003 2.9751e+003PSNR 12.6648 13.3957





Different types of image are taken for discussing theirresults. In our study four different types of images aretaken:a. Face image

Encode Process Time as Second of red =95.9850Decode Process Time as Second of red =13.5620Encode Process Time as Second of green =98.1410Decode Process Time as Second of green =13.7970Encode Process Time as Second of blue =96.9690Decode Process Time as Second of blue =13.6560Total encode time =291.0950Total decode time =41.0150Cr =6.2036Encode_time =291.095secondsThe MSE performance is 56.72 dBThe psnr performance is 30.59 dB

Fig. 6 Face Imageb. Scene image:

Encode Process Time as Second of red =97.8600Decode Process Time as Second of red =13.5630Encode Process Time as Second of green =95.1560Decode Process Time as Second of green =13.4220Encode Process Time as Second of blue =95.9370Decode Process Time as Second of blue =13.4690Total encode time =288.9530Total decode time =40.4540Cr =1.0033Encode time =288.953secondsThe MSE performance is 10.65 dBThe psnr performance is 37.86 dB

Fig. 7 Scene imageV. CONCLUSIONS

Images play an important role in our lives. They areused in many applications. In our study we have applieddifferent type of compression technique on differenttype of images for a PSNR value and compression ratio.In this paper, the problem of compressing an image inMatlab environment has been taken. A matlabcompression approach has been implemented forcompressing images.After analysis we have found that, scene ImagesWavelet provides the better result but compression ratiois high in case of BTC and visual quality is better ofBTC. For face Images BTC perform the mostcompression. FIC provide almost same result ascompare to BTC, but fractal image compression take along time for compression, so we can use BTC as

compare to FIC when time in main concern. Weanalyzed the PSNR obtained of compressed after eachcompression technique and decides which technique canprovide maximum PSNR for a particular image.We have done PSNR measures, but should also useobjective and subjective picture quality measures. Theobjective measures such as PSNR and MSE do notcorrelate well with subjective quality measures.Therefore, we should use PQS as an objective measurethat has good correlation to subjective measurements.After this we will have an optimal system having bestcompression ratio with best image quality.

REFERENCES[1]. A. Kumar, P. Singh, “Comparative study of Block Truncation

Coding with EBTC”, International Journal of Advances inComputer Networks and its Security, pp: 395-405, 2011.

[2]. D. Mohammed, F. Abou-Chadi, “Image Compression UsingBlock Truncating Code”, JSAT, pp 9 – 1, 2011.

[3]. Q. Guan, “Research of Image Compression Based onEmbedded Zero-tree Wavelet Transform”, IEEE, pp: 591-595,2011.

[4]. S. A. Abdul, k. B. A. Karim,” Wavelet transform and fastFourier transform for signal compression: a comparativestudy”, 2011 international conference on electronic devices,systems and applications, IEEE, pp: 280-285, 2011.

[5]. P. aggarwal, B. Rani, “Performance Comparison of ImageCompression Using Wavelets”, IJCSC, IEEE, pp: 97-100,2010.

[6]. Chunlei Jiang, Shuxin Yin,”A Hybrid Image CompressionAlgorithm Based on Human Visual System” InternationalConference on Computer Application and System Modeling,IEEE, pp: 170-173, 2010.

[7]. L. Hengjian, W. Jizhi, W. Yinglong, T. Min, X. Shujiang, “Aflexible and secure image compression coding algorithm”,International Conference on Future Information Technologyand Management Engineering, IEEE, pp: 376-379, 2010.

[8]. S.P. Raja, A. Suruliandi, “Analysis of Efficient Wavelet basedImage Compression Techniques”, Second Internationalconference on Computing, Communication and NetworkingTechnologies, IEEE, pp: 1-6, 2010.

[9]. Shukla, M. Alwani, A. K. Tiwari, “A survey on lossless imagecompression methods” 2nd international conference oncomputer engineering and technology, IEEE vol 6, pp: 136-141, 2010.

[10]. X. Wu, W. Sun “Data Hiding in Block Truncation Coding”,International Conference on Computational Intelligence andSecurity, IEEE, pp: 406-410, 2010.

[11]. T. Y. Mei, T. J. Bo, “A Study of Image CompressionTechnique Based on Wavelet Transform”, FourthInternational Conference on Genetic and EvolutionaryComputing 2010 IEEE, 2010.

[12]. J. Bhattacharjee, “A Comparative study of Discrete CosineTransformation, Haar Transformation, SimultaneousEncryption and Compression Techniques, InternationalConference on Digital Image Processing”, ICDIP, 2009Proceedings of the International Conference on Digital ImageProcessing, Pages 279-283,2009.

[13]. B. Rani, R. K. Bansal and S. Bansal , “Comparison of JPEGand SPIHT Image Compression Algorithms using ObjectiveQuality Measures” , Multimedia, Signal Processing andCommunication Technologies, IMPACT 2009, IEEE, 2009,Page no 90-93, 2009.

[14]. V. P. Lineswala , J. N. patel, “JPEG image compression andtransmission over wireless channel”, International Conferenceon Advances in Computing, Control, and TelecommunicationTechnologies, IEEE, pp: 643-645, 2009.

[15]. M. Grossber, I. Gladkova, S. Gottipat , M. Rabinowitz, P.Alabi, T. George1, and A.Pacheco, “A Comparative Study ofLossless Compression Algorithms on Multi-Spectral ImagerData”, Data Compression Conference, IEEE, pp:447, 2009.


AUTHORSFirst Author– Abhishek

Thakur: M. Tech. inElectronics andCommunication Engineeringfrom Punjab TechnicalUniversity, MBA inInformation Technology fromSymbiosis Pune, M.H.Bachelor in Engineering (B.E.-

Electronics) from Shivaji University Kolhapur, M.H.Five years of work experience in teaching and one yearof work experience in industry. Area of interest: DigitalImage and Speech Processing, Antenna Design andWireless Communication. International Publication: 7,National Conferences and Publication: 6, BookPublished: 4 (Microprocessor and Assembly LanguageProgramming, Microprocessor and Microcontroller,Digital Communication and Wireless Communication).Working with Indo Global College of EngineeringAbhipur, Mohali, P.B. since 2011.Email address: [email protected]

Second Author – RajeshKumar is working asAssociate Professor at IndoGlobal College ofEngineering, Mohali,Punjab. He is pursuing Ph.Dfrom NIT, Hamirpur, H.P.and has completed hisM.Tech from GNE,Ludhiana, India. He

completed his B.Tech from HCTM, Kaithal, India. Hehas 11 years of academic experience. He has authored

many research papers in reputed International Journals,International and National conferences. His areas ofinterest are VLSI, Microelectronics and Image &Speech Processing.

Third Author – Amandeep Batth: M. Tech. inElectronics and Communication Engineering fromPunjab Technical University, MBA in Human ResourceManagement from Punjab Technical University ,Bachelor in Technology (B-Tech.) from PunjabTechnical University . Six years of work experience inteaching. Area of interest: Antenna Design and WirelessCommunication. International Publication: 1, NationalConferences and Publication: 4. Working with IndoGlobal College of Engineering Abhipur, Mohali, P.B.since 2008.Email address: [email protected]

Forth Author – Jitender Sharma: M. Tech. inElectronics and Communication Engineering fromMullana University, Ambala, Bachelor in Technology(B-Tech.)from Punjab Technical University . Five yearsof work experience in teaching. Area of interest:,Antenna Design and Wireless Communication.International Publication: 1 National Conferences andPublication:6 and Wireless Communication). Workingwith Indo Global college since 2008.E-mail address:[email protected]


IJEEE, Vol. 1, Issue 1 (Jan-Feb 2014) e-ISSN: 1694-2310 | p-ISSN:1694-2426

DESIGN OF SELECTIVE ENCRYPTIONSCHEME USING MATLAB

Abhishek Thakur1, Rajesh Kumar2, Amandeep Bath3, Jitender Sharma4

1,2,3,4Electronics & Communication Department, IGCE, Punjab, [email protected], [email protected]

[email protected], [email protected]

Abstract- This paper gives the idea of recentdevelopments in the field of image security andimprovements in image security. Images are used inmany applications and to provide image security usingimage encryption and authentication. Image encryptiontechniques scramble the pixels of the image anddecrease the correlation among the pixels, such that theencrypted image cannot be accessed by unauthorizeduser. This study proposes a method for encrypting thesender’s messages using new algorithm called Chaoticencryption method. This key will be used for encryptingand decrypting the messages which are transmittedbetween two sides. Chaotic encryption technique is thenew way of cryptography. Emphasizing the imagesecurity, this paper is to design the enhanced securealgorithm which uses Chaotic encryption method toensure improved security and reliability.

Keywords— Cryptography, Cryptography key, RGBimage, LSB, image key.

I. INTRODUCTIONAcross the internet information exchanges and the

storage of data in open networks have created anenvironment in which illegal users can obtain theimportant information. Encryption algorithms have beenused to provide a kind of randomness such that anyunauthorized user cannot obtain the secure data. Onlythe authorized user having the key can obtain theoriginal data. Images are used in many applications suchas medical imaging system, military image databases,cable TV, confidential video conferencing, onlinepersonal photograph album, etc. Hence, reliable, fastand robust security techniques are required to store andtransmit digital images. In open environments, there areseveral security problems associated with the processingand transmission of digital images. Therefore it isnecessary to affirm the integrity and confidentiality ofthe digital image being transmitted. The varioustechniques used in the existing image encryptiontechniques can be classified into three major categories:transposition techniques (position permutation),substitution techniques (value transformation) and thecombination i.e. transposition-substitution technique [1].

The position permutation techniques shuffle the pixelposition within the image itself and usually have lowsecurity because histogram of the shuffled imageremains same. If we talk about the value transformationtechniques, they modify pixel value of the originalimage and have the potential of low computationalcomplexity and low hardware cost. Finally, thecombination performs both position permutation andvalue transformation and has the potential of highsecurity. Another technique for securing digital imagesis based on the use of chaotic functions [2].

II. TECHNIQUES USEDIn recent years, a number of different image

encryption schemes have been proposed based onchaotic maps [3-12], in order to overcome imageencryption problems.E. Cryptography

The cryptography has been proposed to ensure theconfidentiality and authenticity of the message. Theencryption key must be long. Yet, it is difficult toremember it and even storing the key in a database or ina file may be insecure. In addition the protection of theconfidentiality of encryption keys is one of theimportant issues to be dealt with. This issue can beefficiently solved through generating the key beforestarting the process of encryption and decryption, ratherthan storing it. Here we will use an image to generatethe key. The main objective of this study is to increasesecurity in communication by encrypting theinformation using a key that is created through using animage. Cryptography presents various methods fortaking legible, readable data, and transforming it intounreadable data for the purpose of secure transmission,and then using a key to transform it back into readabledata when it reaches its destination [8]. Cryptography isconsidered to be one of the fundamental building blocksof computer security [4]. The need of reliable andeffective security mechanisms to protect informationsystems is increasing due to the rising magnitude ofidentity theft in our society. Although cryptography is apowerful tool to achieve information security, thesecurity of cryptosystems relies on the fact thatcryptographic keys are secret and known only to the


legitimate user [9]. In secure communication, keygeneration phase has many challenges and this problemcan be solved if the sender and the receiver share thekey in any other form or if they generate the keysreadily during encryption and decryption separately,thus, the concept of generating the key from an imagecame to the role.F. Visual Cryptography

Visual Cryptography is a special encryptiontechnique to hide information in images in such a waythat it can be decrypted by the human vision if thecorrect key image is used. In secret sharing mechanisms,the secret data is divided into several shares (images)and distributed among participants. Shares are stackedtogether to recover the data. Visual cryptography (VC)attempts to recover a secret image via the human visualsystem by stacking two or more transparencies. In VCapproach [1], the secret was partitioned into n shadowimages (shares), and each participant would receive onlyone share. Once any k or more shares of a secret arestacked together, the secret image will be visuallyretrieved without the help of the computer. That is to saythat the secret image will be invisible if the number ofstacked shares is less than k. This is known as (k, n)threshold mechanism. If the random looking shares areenveloped into some meaningful images the interest ofhackers can be reduced.G. Steganography

Hiding information by embedding secret data intoan innocuous medium is often referred to assteganography. Steganography can be appliedelectronically by taking a message (a binary file) andsome sort of cover (often a sound or image file) andcombining both to obtain a “stego-object”. The RSanalysis is considered as one of the most famoussteganalysis algorithm which has the potential to detectthe hidden message by the statistic analysis of pixelvalues [1]. The process of RS steganalysis uses theregular and singular groups as the considerations inorder to estimate the correlation of pixels [2]. Thepresence of robust correlation has been witness in theadjacent pixels. But unfortunately using traditional LSBreplacing steganography [3], the system enders thealteration in the proportion in singular and regulargroups which exposes the presence of thesteganography. Ultimately, it will not be so hard todecrypt the secret message. Although there are extensiveresearches based on combining these two approaches [4][5] [6], but the results are not so satisfactory withrespect to RS analysis. Other conventional methods ofimage security has witnessed the use of digitalwatermarking extensively, which embeds another imageinside an image, and then using it as a secret image [7].The use of steganography in combination visualcryptography is a sturdy model and adds a lot ofchallenges to identifying such hidden and encrypteddata. The image is divided into a number of blocks andeach block is computed to produce hash value. In thescheme, only the hash value is encrypted using thechaotic logistic map and remaining block values arecomputed using the transition function of linear memorycellular automata. The method uses the concepts oflinear memory cellular automata (LMCA). The linear

memory cellular automata requires large number ofcomputations and one-by-one, each block of the imageis computed using cells of the cellular automata. If themethod is applied over large sized images, more numberof blocks will be made and the number of computationswill be much more. In the scheme, the number of blocksis not constant and can vary. Therefore, if the blocks areof large size, then local transition function will requiremuch computations and if the block size is small thenmore number of blocks will be there, which alsorequires more computations to be performed. Hence, themethod is good butif we apply the scheme over smallcomputation devices (e.g. Micro devices that have lesscomputational power compared to a computer), thenumber of executions will be a complex issue. Thesecurity provided by the scheme is keen but thecomputation needs to be enhanced such that the methodcan be adopted easily.

III. SELECTIVE ENCRYPTIONThe encryption process requires an encryption

algorithm and a key. The process of recovering plaintextfrom cipher text is called decryption. The accepted viewamong professional cryptographers (formalized inKIRKHOFF's law) is that the encryption algorithmshould be published, whereas the key must be keptsecret. Selective encryption scheme generate the keybefore starting the process of encryption and decryption,rather than storing it. In secure communication, keygeneration phase has many challenges and this problemcan be solved if the sender and the receiver share thekey in any other form or if they generate the keysreadily during encryption and decryption separately.

Fig. 1: Encryption of an image.

From Fig. 1, it is clear that the receiver shoulddecrypt the information before it can decompress theimage. This approach has the main drawback that it isimpossible to access the smallest part of informationwithout knowledge of the key. For example, it wouldimpossible to search through a general database of fullyencrypted images. A way to address this issue is to use atechnique called selective encryption; it is depicted in


Fig. 2. The image is first compressed (if needed).Afterwards the algorithm only encrypts part of thebitstream with a well-proven ciphering technique;incidentally a message (a watermark) can be addedduring this process. To guarantee a full compatibilitywith any decoder, the bitstream should only be altered atplaces where it does not compromise the compliance tothe original format. This principle is sometimes referredto as format compliance. WEN et al. [] have recentlydescribed a general framework for format-compliantencryption.

Fig. 2: Selective encryption mechanism.

H. Input ImageThe simplest way to hide binary data on an image is

to use a lossless image format (such as a Bitmap) andreplace the least significant bits of each pixel in scanlines across the image with the binary data. This is notsecure as an attacker can simply repeat the process toquickly recover the hidden information. This technique,known here as “BlindHide” because of the way itblindly hides information, is also not good at hiding –the initial portion of the image is left degraded while therest remains untouched. The proposed project workconsist of mainly two algorithms which are (i)Steganography using Genetic Algorithm (ii) VisualCryptography with Threshold. The application initiateswith Steganography module where the cover image willbe encrypted to generate Stego image. The stagographicimage generated in this module will act as an input forvisual cryptographic module.

Algorithm: SteganographyInput: Cover ImageOutput: Stego Image

11) Read the cover image.12) Find out the pixel values of cover image.13) Read the secret data character wise.14) Convert each character into its equivalent ASCII

code.15) ASCII code is converted into binary values.16) Enter the secret key.17) Secret data is converted into cipher data.18) The stream of 8 bit(cipher data) are embedded into

LSB of each pixel of the cover image.

19) To apply Genetic Algorithm in the stego image thepixel location should be modified.

Algorithm: Visual CryptographyInput: Stego-ImageOutput: Encrypted Shares

1) Read Stego-Image generated.2) The stego image is breaked into three layersnamely split-1, split-2, split-3 these three files arecontaining the hidden data and to get the hidden datathese three files have to be reconstructed perfectly.3) The re-assembled picture and the extracted datawill be gained again. The proposed scheme is based onstandard visual cryptography as well as visual secretsharing. The implementation of the algorithm yields inbetter result with insignificant shares when stego imagesare normally with light contrast. It can also be seen thatthe algorithm gives much darker shares in gray outputthe proposed scheme is based on standard visualcryptography as well as visual secret sharing. Theimplementation of the algorithm yields in better resultwith insignificant shares when stego images arenormally with light contrast. It can also be seen that thealgorithm gives much darker shares in gray output. Thisalgorithm gives better results in terms of image qualityand stegnalysis. good at hiding – the initial portion ofthe image is left degraded while the rest remainsuntouched. The proposed project work consist of mainlytwo algorithms which are (i) Steganography usingGenetic Algorithm (ii) Visual Cryptography withThreshold. The application initiates with Steganographymodule where the cover image will be encrypted togenerate Stego image. The stagographic imagegenerated in this module will act as an input for visualcryptographic module. The original message isembedded into the image by using LSB insertionmethod. The resultant image is called as stego imageshown in Figure 3. Then apply genetic algorithm tomodify the pixel location and detection of message iscomplex.

Then apply visual cryptography scheme stego imageis splitted into two shares based on threshold. The sharesof the stego image is shown in Figure 4.

It is almost impossible for anyone who will attempt todecrypt the encrypted data within that image to reveal ifthe secret shares which they posses are set of allencrypted shares or certain secret shares are missing.

IV. PERFORMANCE ANALYSISThe performance of the proposed system is

experimented by performing stegnalysis and conductingbenchmarking test for analysing parameters like MeanSquared Error (MSE) and Peak Signal to Noise Ratio.Cover image : rice.pngSize : 256*256Mean Square Error (MSE) : 0.0678Peak Signal-to-Noise Ratio (PSNR) : 59.8188dbAfter applying Genetic Algorithm the measuredperformance is shown in belowMean Square Error (MSE) : 0.794Peak Signal-to-Noise Ratio (PSNR) : 39.4011dbAfter applying genetic algorithm all the pixel locationare altered. Due to the change the pixel location MSEand PSNR values are increased.


From Figure 1 and 2, we can conclude that both thesender and receiver before they start to communicatewith each other, they should use the same imagesdatabase, image channel (R, G or B), K value for thekey array, and the session type. This study implementedthe key generation as 1 image/session. The sessionperiod can be an hour or one day. For instance, hourlysession uses one image for each hour and the dailysession uses one image for each day.Phase 1: Database creation:

In this phase, we create a database of color RGBimages which would be used for generating the key forencryption/decryption process. In case of using hourlysession, the database should contain 24 images;otherwise, it should contain 7 images for the dailysession. Both the sender and the receiver should use thesame database of images (24 or 7 images) as the namesof the images should be the same in both sides.I. Key Generation:

In this section, we present the pseudo code for thekey generation algorithm and the algorithm steps asfollows:

1. Get the value of K variable and the value of N frombinaries message step.

2. Select color image from the database according to thesession type and current date or time.

3. Retrievethe image and create array Img [x, y] as LSBof the pixels for one channel (red, green or blue)where the image has x*y dimension.

4. Create array Key [N, k] to represent the encryptionkey.5. Loop i=0 to N-1 /* for each row in Key*/6. Loop c=0 to k/2-1 step 2//fill array by two each7.

Scan the LSB Img[x,y] to find two neighbor valueswhere the absolute difference between them is one.

8. Key[i,c] = First LSB value.9. Key[i,c+1] = Second LSB value.10. End loop11. End loopJ. Image Encryption

In this section, we present the pseudo code for theencryption algorithm and the algorithm steps as follows:1. Get the value of K variable.2. Read a message to encrypt3. Binaries the message and store the binary data in

array Data[N,N] where message bits <= N*N4. Generate the key using key generation algorithm and

create Key [N, k].5. Create array EncData [N, N] to store the encrypteddata6. c=07. Loop i=0 to N-1// for each row in Data /* the

elements in the Data array row i compare with twocolumns of Key array start from c */

8. Loop j=0 to N-1// for each bit in the row i9. If Data[i,j] = Key[j,c]10. EncData[i,j]=0 // the value in the first column11. Else12. EncData[i,j]=1 // the value in the second column13. End loop14. c=c+2// go to next two columns of Key array.

15. if c >= K then c=0// return to the first two columnsof Key array

16. End loop17. Convert the Enc. Data from binary data to text.K. Decryption:

In this section, we present the pseudo code for thedecryption algorithm and the algorithm steps as follows:

1. Get the value of K variable.2. Binaries the encrypted message and store the

binary data in array EncData[N,N]3. Generate the key using key generation algorithm

and create Key [N, k].4. Create Data[N, N] to store the original data5. c=06. Loop i=0 to N-1 /* for each row in Data*//* the

elements in the Data array row i compare with twocolumns of Key array start from c */

7. Loop j=0 to N-1 // for each bit in the row i8. If Encdata[i,j]=0 then

V. RESULTSL. Results for image encryption and decryption

This figure shows the result of image encryption,decryption and histogram of each image.

Fig. 3: Results for image encryption and decryption

Image encryption and decryption algorithm isdeveloped in matlab. In the first step we take any imagewhich we want to encrypt and apply any key consist ofnumber, words and special character. When we applythis key to image its visualization is hidden and itshistogram is also change. When we want to decrypt thisimage we need to enter same key. If we provide wrongkey then image will not decrypt.

Input Image

Encrypted Image

Final Image

0

1000

2000

Histogram of input image

0 100 200

0

2000

Histogram of Encrypted image

0 100 200

0

1000

2000

Histogram of final Image

0 100 200


Fig. 4: Results for image encryption and decryption

VI. CONCLUSIONSImages play an important role in our lives. They are

used in many applications. Therefore it is necessary toaffirm the integrity and confidentiality of the digitalimage being transmitted. Some image encryption andauthentication schemes are discussed in this paper andfurther improvements are also suggested in the previoussection. The chaos based image cryptography isincreasing and improving day by day. The paperpresents different chaos based image cryptographictechniques proposed in the last decade. We discussedthe cryptanalysis of an image encryption scheme andconclude that if the algorithm is not designed properly,it may be possible that the image being protected will beinsecure. Therefore proper security mechanisms shouldbe adopted to secure the image. Each technique isunique and is suitable for different applications.Everyday new image encryption techniques areevolving, hence fast and secure image securitytechniques will always work with high rate of security.We are doing research work in the field of imagesecurity. Image security techniques should be such thatany unauthorized user cannot access the image and ifany modifications can be made to the image, thatmodifications can be detected at the receiving side.Hashing is a good technique that can be used to verifyintegrity of the image. Along with hashing, if properencryption is provided to the image, the image will bemore secure. Using encryption, any unauthorized usercannot access the image. Hence, to improve the image

security, proper combination of encryption as well asauthentication techniques presents ideal alternative. Toachieve high level of image security, we are functioningon a hybrid technique based on image encryption andauthentication. The technique consists of hashgeneration of the plain image. The generated hash isembedded in the image and the entire image isencrypted using combination of chaotic maps andsubstitution-diffusion techniques. The technique canprovide high level security to the image with lesscomputation overheads.

REFERENCES[20] Marc Van Droogenbroeck and Raphael Benedett, “Techniques

for a selective encryption of uncompressed and compressedimages” In Advanced Concepts for Intelligent Vision Systems(ACIVS), Ghent, Belgium, pages 90-97, September 2002.

[21] Amitesh Singh Rajput, "Towards the Growth of ImageEncryption and Authentication Schemes, "2013 InternationalConference on Advances in Computing, Communications andInformatics, P.N.454-459.

[22] Mrs.G.Prema, S.Natarajan, "Steganography using GeneticAlgorithm along with Visual Cryptography for WirelessNetwork Application"

[23] Tawfiq S. Barhoom, Zakaria M. Abusilmiyeh, "A NovelCryptography Method Based on Image for Key Generation", In"2013 Palestinian International Conference on Information andCommunication Technology", 2013 IEEE, P.N. 71-76.

[24] M. Sabery.K, M.Yaghoobi, “A New Approach for Imageencryption using Chaotic logistic map”, 978-0-7695-3489-3/08© 2008 IEEE.

[25] [8] Chen G, Mao Y and Chui CK. “A symmetric imageencryption scheme based on 3D chaotic cat maps”. ChaosSolitons Fractals 2004; 21:749–61.

[26] [9] Zhang Q, Guo L and Wei X. “Image encryption using DNAaddition combining with chaotic maps”. Math Comput Model2010; 52:2028–35.

[27] [10] Sun F, Liu S, Li Z, and Lu Z. “A novel image encryptionscheme based on spatial chaos map”. Chaos Solitons Fractals2008; 38:631–40.

[28] [11] Rhouma Rhouma and Safya Belghith, “Cryptanalysis of anew image encryption algorithm based on hyper-chaos”,Physics Letters A 372(2008) 5973–5978.

[29] [12] Atieh Bakhshandeh and Ziba Eslami , “An authenticatedimage encryption scheme based on chaotic maps and memorycellular automata”, Optics and Lasers in Engineering, Volume51, Issue 6, June 2013, Pages 665-673.

[30] [13] Aloka Sinha and Kehar Singh, “A technique for imageencryption using digital signature”, Optics Communications,Volume, 1 April 2003, Pages 229-234.

[31] [14] Jing Qiu and Ping Wang, “ Image encryption andauthentication scheme”, IEEE, Computational Intelligence andSecurity (CIS), 2011 Seventh International Conference, 3-4Dec. 2011, 784 – 787.

[32] [15] Siddharth Malik, Anjali Sardana and Jaya, “A KeylessApproach to Image Encryption”, IEEE 2012 InternationalConference on Communication Systems and NetworkTechnologies (2012), 879-883.

[14] Narendra K. pareek, Vinod Patidar and K. K. Sud, “Diffusion-Substitution bsed gray image encryption scheme”, DigitalSignal Processing, Volume 23, Issue 3, May 2013, pages 894-901.

AUTHORSFirst Author– Abhishek

Thakur: M. Tech. inElectronics andCommunication Engineeringfrom Punjab TechnicalUniversity, MBA inInformation Technology fromSymbiosis Pune, M.H.Bachelor in Engineering (B.E.-


Electronics) from Shivaji University Kolhapur, M.H.Five years of work experience in teaching and one yearof work experience in industry. Area of interest: DigitalImage and Speech Processing, Antenna Design andWireless Communication. International Publication: 7,National Conferences and Publication: 6, BookPublished: 4 (Microprocessor and Assembly LanguageProgramming, Microprocessor and Microcontroller,Digital Communication and Wireless Communication).Working with Indo Global College of EngineeringAbhipur, Mohali, P.B. since 2011.Email address: [email protected]

Second Author – RajeshKumar is working asAssociate Professor at IndoGlobal College ofEngineering, Mohali,Punjab. He is pursuing Ph.Dfrom NIT, Hamirpur, H.P.and has completed hisM.Tech from GNE,Ludhiana, India. He

completed his B.Tech from HCTM, Kaithal, India. Hehas 11 years of academic experience. He has authoredmany research papers in reputed International Journals,International and National conferences. His areas ofinterest are VLSI, Microelectronics and Image &Speech Processing.

Email address: [email protected]

Third Author – AmandeepBatth: M. Tech. in Electronicsand Communication

Engineering from Punjab Technical University, MBA inHuman Resource Management from Punjab TechnicalUniversity , Bachelor in Technology (B-Tech.) fromPunjab Technical University . Six years of workexperience in teaching. Area of interest: Antenna Designand Wireless Communication. International Publication:1, National Conferences and Publication: 4. Workingwith Indo Global College of Engineering Abhipur,Mohali, P.B. since 2008.Email address: [email protected]

Forth Author – Jitender Sharma: M. Tech. inElectronics and Communication Engineering fromMullana University, Ambala, Bachelor in Technology(B-Tech.)from Punjab Technical University . Five yearsof work experience in teaching. Area of interest:,Antenna Design and Wireless Communication.International Publication: 1 National Conferences andPublication:6 and Wireless Communication). Workingwith Indo Global college since 2008.E-mail address:[email protected]



STUDY OF DIFFERENT TECHNIQUESTO DESIGN ULTRA WIDEBAND

ANTENNAAmandeep Bath1, Abhishek Thakur2, Jitender Sharma3, 4Prof. Basudeo Prasad

1,2,3,4Electronics & Communication Department, IGCE, Punjab, [email protected], [email protected], [email protected]

Abstract—Various papers on differentconfigurations of wideband antenna have been studiedand thereby the different antenna parameters have beenverified. The micro strip antenna has the advantages oflow profile, light weight and low cost. In this paper wewill design a GPS micro strip antenna that perform wideband CP radiation. The proposed antenna consists oforbicular micro strip patch, a substrate of feedingnetwork and a circular patch above and orbicular patchantenna fed by two coaxial is considered as the basestructure. The feeding network was designed andcombined with the antenna. The proposed CP antennacan be designed by adjusting the dimensions of the innerand the outer radius of the orbicular patch to determinethe operating frequency. Then the parameters of thebranch line directional coupler can be fine tuned toachieve the required frequency with good impedancematching.

Keywords- Ultra wide band (UWB); wireless areanetwork (WLAN); circular polarized (CP); GroundPenetrating Radar (GPR); cavity backed Rectangularaperture antennas (CB-RAA); transmitter and receiver(T/R).

I. INTRODUCTIONCircular polarization antennas are needed due to

their insensitivity to ionospheric polarization rotation.Besides, for terminals to receive satellite signals, a wideradiation beam is extremely needed. Micro strip antennahas been widely used due to its small size, lightweight,low profile, low cost and easy to conform. Agilereconfigurable antennas for future communicationsystems have attracted researchers around the globe.Antenna's characteristics such as frequency, radiationpattern and polarization are reconfigured to attain thedemands for agile radios. A lot of researches focus onfrequency reconfiguration as future communicationsystems such as cognitive radio needs an antenna thatcan do spectrum sensing and communication. Inreconfigurable frequency antennas development,recently a reconfigurable wide band to agile narrowfrequencies, using a printed log periodic dipole arrayantenna, was introduced. A wideband slottedmultifunctional reconfigurable frequency antenna forWLAN, WIMAX, UWB and UMTS has been proposedin, a frequency reconfigurable antenna, consisting oftwo structures; one is an ultra wide band (UWB) and

other is a frequency reconfigurable triangle shapeantenna, is proposed for cognitive radio communication.

Microstrip antennas have been widely used in manymodern communication systems, because of itsrobustness, planar profile, and low cost. Most of theseantennas operate at their fundamental mode, whichgives a broadside beam. Microstrip antenna operating atthe higher order mode has dual symmetric radiationbeams. Microstrip antennas are made up of a very thinmetallic strip, which is a patch on a grounded substrate,found numerous applications in different fields due totheir wonderful characteristics. These antennas areavailable with a very low profile, they are light inweight, very compact and conformable structure andeasy to be fabricated. These antennas have drawnattentions of scientific community over the past decadesseveral reconfigurable have been designed for cognitiveradio applications. An effective method for doing this isto use physical alteration of the antenna parts via arotational motion to achieve the reconfigurablecharacteristics. Antennas used in ultra wide band(UWB) systems have been extensively studied in therecent years. Ultra wide band antennas have alreadybeen used in areas such as satellite communication,remote sensing, ultra wide band radar and so on.Currently, the wireless area network (WLAN) in the2.4-GHz (2.4-2.485 GHz) and 5-GHz (5.l5-5.875 GHz)bands is the most popular networks for accessing theinternet the antenna for an AP not only requires dualband operation but also needs to have an appropriateradiation profile in both bands, namely similar gain,wide beam width, and high front-to-back ratio. Thefuture generation wireless networks require systemswith broad band capabilities in various environments tosatisfy several applications as smart grid, personalcommunications, home, car, and office networking [1].

The attractive advantages of the CP antenna areexisted as follows. Firstly, since the CP antennas sendand receive in all planes, it is strong for the reflectionand absorption of the radio signal. In the multi pathfading channel environment, the CP antenna overcomesout of phase problem which can cause dead spotsdecreased throughput, reduced overall systemperformance. Additionally, The CP antenna is moreresistant to signal degradation due to inclement weatherconditions. Circular polarization (CP) is commonlyadopted in GPS and other satellite communications dueto the phenomenon of Faraday rotation when signalstravel through the ionosphere. The circularly polarized


wave can be realized by exciting two linearly polarizedmodes. These two modes should be with 90 degreephase difference, equal amplitude, and orthogonal toeach other in polarization. There are many types ofantenna that can carry out the CP wave, such asspiral[2].

But the micro strip antennas would be our firstchoice. The proposed antenna consists of orbicularmicro strip patch, a substrate of feeding network and acircular patch above and orbicular patch antenna fed bytwo coaxial is considered as the base structure. Ultrawide band (UWB) impulse technologies are being usedincreasingly for high speed RF wireless communication,high power RF jamming and high resolution impulseradar systems. The UWB operation provides criticaladvantages, such as improved detection, ranging andtarget resolution performances (UWB) GPR system thattransmits short time impulse signal is used to benefitfrom both low and high frequencies. For the best UWBperformance, the transmitter and receiver (T/R) antennasshould have flat and high directive gain, narrow beam,low side and back lobes over the operational frequencyband; to attain the largest dynamic range, best focusedillumination area, lowest T/R coupling, reduced ringingand uniformly shaped impulse radiation. UWB haspromised to offer high data rates at short distances withlow power, primarily due to wide resolution bandwidth[3].

II. DIFFERENT TECHNIQUES USEDWide band phased array antenna is implemented to

scan in the range of +-45 degree in wideband. A novelwideband mono conical antenna in the shape of tearsdrop is implemented with good radiation directivity andfrequency response. A dual spiral antenna for ultrawideband capsule endoscope system is implemented.The capsule endoscope system transmits the image dataof internal human body to an external receiverwirelessly. This system contains sensors, a tiny camera,LEDs, a wireless IC transceiver and antenna and theseare encased in small capsule. The antenna has omnidirectional radiation pattern to transmit signalsindependent of capsule position and orientation. Awideband circularly polarized planar monopole antennais implemented. It features the simple feeding networksthat are composed of a Wilkinson power divider and 90◦phase shifters. However, impedance bandwidth isaround (20% to 30%). The attractive characteristics ofthe proposed antenna are the wide impedancebandwidth of 87.3% (1 GHz to 2.55 GHz), the 3 dBaxial ratio bandwidth of 95% (1.05GHz to 2.95 GHz).Wideband Circularly polarized micro strip antenna isimplemented for global positioning systems at 1575 Hzband wide beam circular polarization micro stripantenna with bandwidth from 3.46 GHz to 4.36 GHz.There are many types of antenna that can carry out theCP wave, such as spiral [4].

But the micro strip antennas would be our firstchoice as it has the advantages of low-profile,lightweight and low cost. In this paper we will design aGPS micro strip antenna that perform wide-band CPradiation. Such antenna can be actualized using a singlefeed or dual feeds. The single-feed CP antenna is simple

in structure and easy to fabricate, while the bandwidth israther to narrow. Therefore in this letter dual feeds areadopted to realize the wide-band CP antenna for dual-band GPS applications. The circularly polarized wavecan be realized by exciting two linearly polarizedmodes. These two modes should be with 90 degreephase difference, equal amplitude, and orthogonal toeach other in polarization. Such antenna can beactualized using a single feed or dual feeds. WidebandU slot micro strip antenna operating in the frequencyrange of 5.18 to 5.8 GHz and improving directivity.Two radiation beams off broadside are obtained byoperating the patch antenna at the higher order modeinstead of the fundamental mode, which radiates abroadside beam. Broadening the antenna bandwidth isachieved by using the U slot technique. Micro stripantennas have been widely used in moderncommunication systems, because of its robustness,planar profile, and low cost. Most of these antennasoperate at their fundamental mode, which gives abroadside beam. Micro strip antenna operating at thehigher order mode has dual symmetric radiation beams[2]. It is well known that the major drawback of a microstrip antenna is its narrow bandwidth (3%). One of thepopular techniques for broadening the patch antennabandwidth is to incorporate a U slot on its surface asproposed [5].

Ultra wideband millimeter wave stack patchantenna achieving high efficiency >90% and high gainof 6 to 8 db. Antenna is fabricated on silicon and takesadvantage of cavity etching and BCB membranesfabrication process of an UWB 60-GHz antenna basedon silicon micromachining and BCB(BenzoCycloButene) membranes to achieve highefficiency and large band performances the use ofmembranes to reduce the effective permittivity aroundthe radiating elements in order to reduce the effectivepermittivity below the top patch, a 280- μm cavity wasetched in the cap [6] .

Implementation of wideband antenna for smart gridapplications with a frequency bandwidth of 40% andgain of 3 to 4db.The design and the required simulationof the antenna is done using ANSYS’ HFSS softwarewhich is the standardized simulation tool for 3dimensional full-wave electromagnetic field simulation.The paper gives the design of a novel compact lowprofile wide bandwidth micro strip antenna. The totalsize of the antenna is 20mm x 10mm x 2mm. Thisavailable gain from this new design is from 3db to 4.3db and it offers a wide fractional frequency bandwidthwhich is around 40% over the frequency band (5GHz –7.5GHz) [7].

Using ultra wideband dipole antenna operating at1.75 to 40 GHz .It is shown that the proposed antennaworks well in 1.7GHz-40GHz frequency range and themain direction of the radiation pattern keeps stableduring the whole frequency range. The H planedemonstrates an excellent omni directional pattern aDual band Wide beam width WLAN Access PointAntenna with similar gain and wide beam width in boththe 2.4- and 5-GHz WLAN bands. This paper describesa dual-band printed dipole antenna that has nearlyidentical radiation patterns with similar gain and wide


beam width in both the 2.4-GHz and 5-GHz WLANbands. The proposed design employs two techniques toimprove the radiation pattern. These techniques are theuse of an angle dipole and vertical copper platesarranged on the ground plane for improvement in theradiation pattern of lower and upper bands, respectively.Ultra Wide Band TEM Horn Antenna for GroundPenetrating Radar (GPR) is used and broad band widebeam circular polarization micro strip antenna withbandwidth from 3.46 GHz to 4.36 GHz. Since TEMhorn is a kind of travelling wave antenna, its structurecan be considered as combination of micro-striptransmission line segments, which are characterized bytheir local geometrical and constitutional structureparameters. The staircase modeling is used for theanalysis. In this work, different design forms of TEMhorn antennas such as, dielectric loaded, Vivaldi shapedand array versions were surveyed for UWB GPRsystems [8].

Circular polarization is used which is produced byexciting two linearly polarized modes of antenna with90 degree phase difference, equal amplitude andorthogonal to each other. To improve the performanceof antenna the operating frequency of the array antennais varied and the electrical performance and the couplingbetween the elements will be different [9]. Varioustechniques have been explored to develop small sizeantenna. One of the techniques is achieved by usingshorting wall another technique is the shortening of theground plane for bandwidth enhancement and sizereduction. The design has a ground plane and a metalcircular disk which is kept perpendicular to the groundplane giving omni directional characteristics directiveultra wideband rectangular waveguide based antenna.To overcome these limitations, we investigate the use ofcavity backed Rectangular aperture antennas (CB-RAA)to achieve very good radiation characteristics andexcite the fundamental mode over very wide bandwidth.

Fig. 1: Antenna structure: (a) isometric view [4] (b) front view[4]

The antenna for an AP not only requires dual bandoperation but also needs to have an appropriateproposed design. It employs two techniques to improvethe radiation pattern. These techniques are the use of anangled dipole and vertical copper plates arranged on theground plane for improvement in the radiation pattern oflower and upper bands, respectively radiation profile inboth bands, namely similar gain, wide beamwidth. Oneof the popular techniques for broadening the patchantenna bandwidth is to incorporate a U-slot on itssurface. They have the ability to confine the power incertain directions instead of scattering the powereverywhere. As a result of less power loss toward

unwanted directions, the multipath and interferenceeffects are reduced.Using wideband circularly polarized monopole antenna.The wideband circularly polarized planar monopole[10]. The proposed antenna is formed by four simplewideband planar monopole antenna elements in order toimprove the performance of circular polarization itsfeatures the simple feeding networks that are composedof a Wilkinson power divider and 90◦ phase shifters;however impedance bandwidth is around (20 ∼ 30)%.The attractive characteristics of the proposed antennaare the wide impedance bandwidth of 87.3% (1 GHz to2.55 GHz), the 3 dB axial ratio bandwidth of 95% (1.05GHz to 2.95 GHz).

Fig. 2: Proposed antenna prototype [6]

Fig. 3: Photograph of the fabricated proposed antennaprototype: Bottom view of Novel Ice cream cone ultra

wideband antenna.[9]

It is challenging to enhance their bandwidth. Toaccount for this a novel antenna shape named an IceCream Cone UWB antenna has been designed tonumerous UWB applications like wirelesscommunication meet the demands of a reduced antennasize, a higher dielectric indoor positioning and medicalimaging. The printed planar substrate is required forimplementing mono conical antenna for widebandelectromagnetic field generation. A novel widebandantenna of which shape is tears drop type is proposed byevolving the antenna combined a hemisphere with a biconical antenna element.

Fig. 4: shows the structure of the antenna that combineda sphere with a mono-conical antenna element

smoothly. The radius of the spherical part in the antennais 125 mm, and the antenna element is made by styrene

foam covered with cupper tapes [14].


The antenna is composed a mono conical antennaelement, which is set on a metal floor in an anechoicchamber and can generate the high powerelectromagnetic field above 1GHz. Fundamentalcharacteristics of this antenna such as the radiationdirectivity and the frequency response are measured.Figure shows a picture of the proposed antenna that hasa shape of “tears drop”. Using Novel Ultra WidebandPlanar Reflector Antenna a reflector antenna relies onreflection of energy from the reflector. Reflectorantennas tend to be good directional and high gainantennas [11]. There are some classes of ultra widebandreflector antennas, such as planar, corner, and paraboliccylinder reflector antennas. In these reflector antennas,the planar reflector antenna is the simplest and yet mostpractical design. Principle of the planar reflector antennais image theory. As shown in figure according toantenna image theory, an image antenna can generate inthe other side of the perfect electric conducting plane.Using conducting plane as a back reflector, a morecompact antenna structure may be feasible. Theconducting planar reflector serves to concentrate theradiation to the antenna side of the planar reflector [12].

Fig. 5: Return loss of novel ultra wideband planarreflector antenna [6].

Fig. 6: .Simulation of radiation patterns [8]

A dual spiral antenna for Ultra wideband capsuleendoscope system a dual spiral antenna for ultrawideband capsule endoscope system is proposed. Sincea capsule endoscope system which transmits real timeimage data in the body should have ultra-wideband

characteristic, an ultra wideband antenna is suitable forthis system. When an antenna is used as capsuleendoscope system, a small sized antenna is requiredbecause capsule is small enough to be swallowed. Forthese two conditions, the proposed antenna has dualresonance structure which is composed of two differentspiral elements and single feed wire connects these twoelements [13]. The bandwidth of the proposed antenna is411V600MHz (189 MHz) for VSWR < 2 which meansthe fractional bandwidth of 37.8% and it has isotropicradiation pattern. Ultra wide band (UWB) impulsetechnologies are being used increasingly for high speedRF wireless communication, high power RF jammingand high resolution impulse radar systems.

Fig. 7: Radiation pattern of circularly polarizedMicro strip antenna [9]

III. COMPARISION OF TECHNIQUESCircular polarization technique; Rotational motion

of antenna technique; Ultra wide band angled dipoleantenna technique; Wideband U slot micro stripantenna technique Variation of the operating frequencytechnique; Dual band access point antenna technique;Ice cream cone ultra wideband technique; Widebandcircularly polarized monopole antenna Mono conicaltears drop type antenna; UWB rectangular waveguidebased antenna [14].

With the variation of operating frequency amongthe above mentioned techniques the wideband phasedarray antenna provides the best scanning range. Theultra wideband antenna fabricated on silicon with cavityetching provides the highest efficiency among all thedifferent techniques used for the improvement ofantenna parameters [15]. Further it provides animprovement in the overall gain of the antenna by 6 to 8db. Further wideband U slot micro strip antennaprovides the broadening of the antenna bandwidth andthe overall improvement in the directivity of theantenna. Ultra band dipole antenna and circularlypolarized antenna provides the best Omni directionalradiation pattern. Also the techniques such as angleddipole and vertical copper plates on ground plane areused for the further improvement of the radiation patternof the antenna. The circularly polarized micro stripantennas are the one providing the highest gain andefficiency [16]. Also the best operating frequency bands,directivity and the antenna radiation pattern. Circularpolarization antennas are needed due to their


insensitivity to ionospheric polarization rotation. Adual-feed wide-band circularly polarized micro stripantenna is presented with an orbicular patch for GPS userequiring approximately 20 MHz at 1227 MHz and 1575MHz each is considered as the one providing best resultsfor improvement of antenna parameters. Also furtherimprovements could be done by using antennasubstrates with higher dielectric constants in order toreduce the size a broad band wide beam circularpolarization micro strip antenna. The configuration ofthe antenna is simple and easy to fabricate comparedwith conventional micro strip antenna, the radiationbeam is broadened obviously. Further research oncircularly polarized wideband micro strip antenna isrequired as it gives the best performance and overallimprovement of antenna parameters [17].

IV. CONCLUSION AND FUTURE WORKWith the rapid progress of wireless technology in

recent years, various wireless systems such as GSM,WCDMA/UMTS, Bluetooth, WLANs, and GPS havebeen highly integrated into the mobile devices, and inorder to fulfill the RF system requirements using thedifferent frequency band, antenna technology isrequired to wideband characteristics. On the other hand,many modern wireless Communication systems such asradar, navigation, satellite,and mobile applications use the circular polarized (CP)

radiation pattern. The attractive advantages of the CPantenna are existed as follows. Firstly, since the CPantennas send and receive in all planes, it is strong forthe reflection and absorption of the radio signal. In themulti-path fading channel environment, the CP antennaovercomes out of phase problem which can cause dead-spots, decreased throughput, reduced overall systemperformance. Additionally, The CP antenna the signaldegradation may be there due to poor weatherconditions. Circular polarization antennas are neededdue to their insensitivity to ionospheric polarizationrotation. A dual-feed wide-band circularly polarizedmicros trip antenna is presented with an orbicular patchfor GPS use requiring approximately 20 MHz at 1227MHz and 1575 MHz each is considered as the oneproviding best results for improvement of antennaparameters [18]. Also further improvements could bedone by using antenna substrates with higher dielectricconstants in order to reduce the size a broad band widebeam circular polarization micro strip antenna. Theconfiguration of the antenna is simple and easy tofabricate compared with conventional micro stripantenna, the radiation beam is broadened obviously.Further research on circularly polarized widebandmicro strip antenna is required as it gives the bestperformance and overall improvement of antennaparameters.

REFERENCES[33] Gaboardi P., Rosa L., Cucinotta A., and Selleri S.,

“Patch Array Antenna for UWB Radar Applications”, in3rdEuropean Radar Conference, 2006, p.281-284.

[34] Yoann Letestu and Ala Sharaiha, “Size reduced multi-band printed quadrifilar helical antenna,” IEEE Trans.Antennas Propag., vol. 59, pp. 3138-3143, 2011.

[35] A. Siligaris et al., “A 65-nm CMOS fully integratedtransceiver module for 60-GHz Wireless HDapplications,” IEEE Journal of Solid-State Circuits, vol.46, no. 12, pp. 3005-3017, Dec. 2011.

[36] S. Manafi, S. Nikmehr, and M. Bemani, "Planarreconfigurable multifunctionalantennaforWLAN/wimax/UWB/pcsdcs/UMTSapplications," Progress In Electromagnetics Research C,Vol. 26, 123- 137, 2012.

[37] C. R. Medeiros, E. B. Lima, 1. R. Costa, and C.A.Fernandes, "Wideband slot antenna for WLANaccesspoint, " IEEE Antenna Wireless Propagate. Lett.,vol. 9,pp. 79-82, 2010.

[38] F. Ghanem, P. S. Hall and J. R. Kelly, “Two portfrequency reconfigurable antenna for cognitive radios”,Electronics Letters,vol.45, 2009,pp.534-536

[39] E. Ebrahimi, J. R. Kelly and P. S. Hall, “Areconfigurable Narrowband antenna integrated withwideband monopole for cognitive radio applications”,IEEE Antennas and Propagation Society InternationalSymposium( APSURSI), 2009.

[40] J. W. Baik, S. Pyo, T.H. Lee, and Y.S. Kim, “Switchableprinted Yagi- Uda antenna with patternreconfiguration”, ETRI Journal, vol.31 2009,pp.318-320

[41] M. Sanad, "A Small Size Microstrip Antenna Circuit",IEEE International Conference on Antenna andPropagation, vol. 1, pp. 465-471, April1995.

[42] P. Suraj and V. R. Gupta, “Analysis of a RectangularMonopole Patch Antenna”’International Journal ofRecent Trends in Engineering,Vol. 2, No. 5, pp. 106-109, November 2009.

[43] M. N. Srifi, M. Meloui and M. Essaaidi, “RectangularSlotted Patch Antenna for 5-6GHz Applications”,International Journal of Microwave and OpticalTechnology, Vol.5 No. 2, pp., 52-57 March 2010.

[44] G. Augustin, S. V. Shynu, C. K. Aanandan, and K.Vasudevan, "Compact dual-band antenna for wirelessaccess point" Electron. Lett., vol. 42, no. 9, pp. 502-503,Apr. 2006.

[45] S. W. Su, "Concurrent dual-band six-loop-antennasystem with wide 3-dB beamwidth radiation for MTMOaccess point" Microwave Opt. Techn. Lett., vol. 52, no.6, pp. 1253-1258, Jun. 2010.

[46] S. W. Su, "High-gain dual-loop antenna for MTMOaccess point in the 2.4/5.2/5.8 GHz bands, " IEEETrans.Antenna Propag., vol. 58, no. 7, pp. 2412-2419,Jul. 2010.

[47] K.B. Hsieh, M.H. Chen, and K.L. Wong, "Single-feedDual-Band Circularly Polarized Microstrip antenna,"Electron. Lett. VOL. 34, pp.1170-1171,Jun.1998.

[48] G. P. Jin, D.L. Zhang and R.L. Li, “Optically controlledReconfigurable antenna for cognitive radiosapplications”, Electronics Letters, vol.47,2011,pp.948-950.

[49] Wang Yazhou, Su Donglin, Xiao Yongxuan,“Broadband circularly polarized square microstripantenna”, Antennas, Propagation and EMTheory, 2006,ISAPE 2006 7th International Symposium on, pp.1–4,2006.

[50] Zhen-Yu Zhang, Yong-Xin Guo, Ling Chuen Ong, ChiaM.Y.W., “A New wide-band planar balun on a single-layer PCB”, Microwave and WirelessComponentsLetters, IEEE, Vol. 15, No. 6, pp. 416–18, Jun. 2005.

AUTHORS


First Author – Amandeep Batth: M. Tech. inElectronics and Communication Engineering fromPunjab Technical University, MBA in Human ResourceManagement from Punjab Technical University ,Bachelor in Technology (B-Tech.)from PunjabTechnical University . Six years of work experience inteaching. Area of interest: Antenna Design and WirelessCommunication. International Publication: 1, NationalConferences and Publication: 4. Working with IndoGlobal College of Engineering Abhipur, Mohali, P.B.since 2008.Email: [email protected]

Second Author– AbhishekThakur: M. Tech. in Electronicsand CommunicationEngineering from PunjabTechnical University, MBA inInformation Technology fromSymbiosis Pune, M.H.Bachelor in Electronics (B.E.)from Shivaji UniversityKolhapur, M.H. Five years ofwork experience in teachingand one year of work

experience in industry. Area of interest: Digital Imageand Speech Processing, Antenna Design and WirelessCommunication. International Publication: 7, NationalConferences and Publication: 6, Book Published: 5(Microprocessor and Assembly Language Programming,Microprocessor and Microcontroller, DigitalCommunication and Wireless Communication).Working with Indo Global College of EngineeringAbhipur, Mohali, P.B. since 2011.Email: [email protected]

Third Author – Jitender Sharma: M. Tech. inElectronics and Communication Engineering fromMullana University, Ambala, Bachelor in Technology(B-Tech.)from Punjab Technical University . Five yearsof work experience in teaching. Area of interest:,Antenna Design and Wireless Communication.International Publication: 1 National Conferences andPublication:6 and Wireless Communication). Workingwith Indo Global college since 2008E-mail: [email protected]

.



BER PERFORMANCE OF MU-MIMOSYSTEM USING DIRTY PAPER CODING

Garima Saini1, Shivkaran Meghwal21,2Electronics and Communication Engineering, National Institute of Technical Teachers, Training and Research,

Chandigarh, [email protected], [email protected]

Abstract- In this paper Dirty Paper Coding forcommunication system is implemented. MIMOapplication that involves devices such as cell phones,pocket PCs require closely spaced antenna, whichsuffers from mutual coupling among antennas and highspatial correlation for signals. DPC is used forcompensating the degradation due to correlation andmutual coupling. Simulation results show significantperformance in terms of bit error rate (BER) by use ofDirty Paper Coding (DPC) for 4G communication.

Index Terms- Dirty Paper Coding (DPC), Multi-UserMIMO (MU-MIMO), Broad-Cast channels (BC),Multi-access channels (MAC), Correlation, Mutualcoupling.

I. INTRODUCTIONWireless is an emerging field, which has beenenormous growth in last several years. The huge uptakerate of mobile phone technology, wireless local areanetwork (WLAN) and exponential growth of theinternet have resulted in an increased demand for newmethods of obtaining high capacity wireless network.The goal of 2G, 3G and 4G is to provide a wider rangeof services like as communications, video phones, andhigh speed internet access. To meet the requirements ofemerging high bandwidth applications, wireless systemscontinue to strive for higher and higher data rates [1].Large spectral efficiencies have been predicted for

wireless system with multiple antennas when thechannel exhibits rich scattering. It has been shown thatMIMO systems have the potential for large informationtheoretic capacities. They provide several independentcommunications channels between transmitter andreceiver. In an ideal multipath channel, the MIMOcapacity is approximately N times the capacity of asingle system, where N is the smaller size of thetransmit or receive antenna elements. The channelcapacity of MIMO system is found to be limited bycorrelation [2]. The spectral efficiency of 3G network istoo low to support high data rate services at low cost.Since as soon as in MIMO 3G system the number ofantenna elements will be increased, due to this capacitywill reduced [3, 4]. As a consequence one of the mainfocuses of 4G is to significantly improve the spectralefficiency. This requirement of improvement in spectralefficiency makes use of Dirty Paper Coding [5]. Forusing application such as Wireless LAN, Cellulartelephony, single base station must communicate withmany users simultaneously. Therefore, the study ofMulti-User MIMO (MU-MIMO) systems has emerged

as an important research topic recently. The channelcapacity of single user NR x NT MIMO systems isproportional to Nmin=min (NT, NR) [6]. In the Figure 1,shows multiple users are connected with station [7].

Figure 1. A Multi-user MIMO system for K=4[7]

Four users are connected into Figure 1. i.e. K=4. Threeout of four users are selected and allocatedcommunication resource such as time, frequency, andspatial stream. In multi-user K.NM antenna cancommunicate with a single BS antenna with NB

antennas. So, (K.NM) x NB system are used fordownlink and NB x (K.NM) MIMO system for uplink.

II. MATHEMATICAL MODEL FOR MULTI-USER MIMO [7]

A. Uplink Channel(Multiple access channel)In Figure 2. Uplink channel is mathematically modeled.We assume that Base stations (BS) and mobile station(MS) are equipped with NB and NM. The received signalis given by zxHxHxHy K

UL

K

ULUL

MAC ............

2211

(1)

z

x

xHHH

K

UL

K

ULUL

1

21...............

x

xH

K

UL

1

(2)


Figure 2. Uplink channel model for MU-MIMO

The downlink model is shown in Figure 3. The receivedsignal is given by

zHy u

DL

uux , where u=1, 2, 3……..K

(3)The overall system can be represented by followingequations

y

yy

K

2

1

H

HH

DL

K

DL

DL

2

1

+

z

zz

K

2

1

(4)

Figure 3. Downlink channel model for MU-MIMO

II. ANALYSIS OF DIRTY PAPER CODINGDirty paper coding is a coding technique that pre-cancels known interference without power penalty.Only the transmitter needs to know this interference,but full channel state information is requiredeverywhere to achieve the weighted sum rate DirtyPaper Coding. Dirty Paper Coding techniquerequires knowledge of the interference state in a non-causal manner [8]. The design of a DPC-based systemshould include a produce to feed side information to thetransmitter. Interference free transmission can berealized by subtracting the potential interferences beforetransmission. The working of Dirty Paper Coding maybe explained by the Figure 4.

Figure 4. Communication system model using Dirty PaperCoding

The received signal for such system is given by

S= Z+P+Q(5)

Where, P is arbitrary interference known at transmitter,N is statistically independent Gaussian random variable.If known interference P is subtracted at receiver, itposes no problem. Similarly, known interferencesubtracts from transmitter, then transmitted signal

Z'=Z-P(6)Now, the received signal is given by

S'=Z'+P+Q(7)

S'=Z-P+P+Q

=Z+Q

(8)

IV. MATHEMATICAL EXPRESSION OF DIRTYPAPER

CODING [7]Let consider the case of NB=4 (Number of Base stationantennas), K = 4 (Number of Users) and NM,u =1(Number of user at Mobile station) where u=1,2,3,4. If

the uth user signal is given by Cx u~ , then the

received signal is given as

zzzz

xxxx

HHHH

yyyy

DL

DL

DL

DL

4

3

2

1

4

3

2

1

4

3

2

1

4

3

2

1

~~~~

(9)

Here CHDL

u

41 is the channel gain between BS

and uth user. The Channel matrix HDL

can be

decomposed as


qqqq

lllllll

lll

HDL

u

4

3

2

1

44434241

333231

2221

11

0

00

000

(10)

Here qqqq4321

,,, are orthonormal vectors. Let

Txxxxx4321

denotes a pre-coded

signal

for Txxxxx ~~~~4321

~ . Transmitting xQH

,

the effect of Q in equations (10) is eliminated through

the channel by leaving the lower triangular matrix aftertransmission. The received signal is given as

zzzz

Q

HHHH

yyyy

xH

DL

DL

DL

DL

4

3

2

1

4

3

2

1

4

3

2

1

zzzz

xxxx

lllllll

lll

4

3

2

1

4

3

2

1

44434241

333231

2221

11

0

00

000

(11)From the equation (11) the received signal for the firstuser is given by

zxly 11111

(12)The equation for interference free for first user is

xx ~11

(13)

By the same process equations for second, third andfourth users are as follows

xllxx ~~

122

2122

(14)

xllxl

lxx 233

321

33

3133

~

(15)

xllxl

lxllxx 3

44

432

44

421

44

4144

~

(16)

The pre-coded signal by the equations (13), (14), (15)and (16) can be expressed as

xxxx

xxxx

~~~~

~~~

4

3

2

1

4

3

2

1

1000

0100

0010

0001

(17)

xxxx

ll

xxxx

~~~

~~

4

3

2

1

22

21

4

3

2

1

1000

0100

001

0001

(18)

xxxx

ll

ll

xxxx

~~

~4

3

2

1

33

32

33

31

4

3

2

1

1000

01

0010

0001

(19)

xxxx

ll

ll

ll

xxxx

~4

3

2

1

44

43

44

42

44

41

4

3

2

1

1

0100

0010

0001

(20)By combining these four equations following pre-codeddirty paper coding is achieved.

zzzz

xxxx

ll

ll

yyyy

4

3

2

1

4

3

2

1

44

33

22

11

4

3

2

1

~~~~

000

000

000

000

(21)It is concluded that Dirty Paper Coding is a scaledinverse matrix of the lower triangular matrix which isobtained from the channel gain matrix.

V. SIMULATIONS AND DESIGN

For simulations, matrix is taken with perfect channelstate information at the receiver. QPSK modulationscheme is used for transmission. Equal power allocation


0 2 4 6 8 10 12 14 16 18 200

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Eb/No (dB)

Bit

Erro

r Rat

e

Bit error rate versus Eb/No

Without DPCWith DPC

is considered for all antennas at the transmitter. Idealantenna length, λ/2 is taken for analysis [9]. Antennaarrays arrangement is side by side system. Thefollowing design specifications are taken.

Table 1.Desisgn specifications for Dirty Paper CodingNO. OF FRAMES 10

NO. OF PACKETS 250

NO. OF BASE STATION (BS)ANTENNA

4

NO. OF MOBILESTATION(MS)

ANTENNA

4

NO. OF USERS 10,20,30

VI. RESULTSBit error rate analysis is done for 4×4 matrix for withDirty Paper Coding and without Dirty Paper Coding.The simulation is done between 0 to 20 dB. Thenumber of iteration is 5000. Figure 5, 6 and 7 show thesimulation results when numbers of users are 10, 20,and 30 at mobile station.

Figure 5. BER when number of users are 10

Figure 6. BER when number of user are 20

Figure 7. BER when number of users are 30Analysis shows that as the numbers of users increase,the performance of Dirty Paper Coding increases. It isobserved that BER is reduced approximately 25% to35% at 6 dB With Dirty Paper Coding.

VII. CONCLUSIONFrom the analysis it is concluded that system has betterbit error performances when Dirty Paper Coding isused. Bit Error Rate is reduced when the numbers ofusers are increased

REFERENCES[1] Z.Tu and R.S. Blum, “ Multi-user diversity for a DirtyPaper approach,” IEEE Communication, Letter , Vol. 7, no. 8,pp. 370-372, 2003.[2] G.J.Foschini and M.J Gans, “On Limits of WirelessCommunications in a Fading Environment when UsingMultiple Antennas,” Wireless Personal Communications, Vol.6, pp. 311-315, 1998.[3] P. N. Fletecher, M. Dean, and A. R. Nix, “Mutualcoupling in multi element array antennas and its influence onMIMO channel capacity,” Electronics Letters, Vol. 39, no. 4,pp. 342-344, 2003.[4] R. Janaswamy, “Effect of element mutualcoupling on the capacity of fixed length linear arrays,”Antennas and Wireless Propagation Letters,Vol. 1, no.8,pp.157-160, 2002.[5] Aditya kumar, Ajith Bhatt, Anil M .V., Prahlad TKulkarni, “Dirty Paper Coding- A Novel Approach forCompact MIMO Systems,” IEEE Second InternationalConference on Communication System, Network andApplication, pp. 86-89, 2010[6] Foschini,G.J, “Layered space–time architecture forwireless communication in fading environment whenusing multi-element antennas,” Bell Labs Tech.J.,1(2), pp. 41-59,1996.[7] Yong Soo Cho, Jaekwon kim Won Young Yang, Chung-Gu Kang, MIMO-OFDM Wireless Communication withMatlab, 2nd ed., Wiley Singapore, 2010.[8] U.Erez, S.Shami, and R.Zamir,“Capacity and latticestrategies for cancelling known interference,” in Proc. Int.Symp. of Information Theory Application , pp. 681-684,2000,

0 2 4 6 8 10 12 14 16 18 200

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Eb/No (dB)

Bit

Erro

r Rat

e


Without DPCWith DPC

0 2 4 6 8 10 12 14 16 18 200

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Eb/No (dB)

Bit

Err

or R

ate


Without DPCWith DPC


[9] C.A.Balanis, Antenna Theory: Analysis and Design, 2nd ed, Wiley, New York, 1997


OPTIMIZATION:MVSIS V/S AIG REWRITING (ABC)

Manish Kumar GoyalDeptt. Of Electronics Engineering, Govt. Polytechnic College, Alwar, Rajasthan, India

[email protected]

Abstract- For the optimization of given network,VHDL /Verilog code convert into BLIF / BLIF_MV(Berkeley Logic Interchange Format /Berkeley LogicInterchange Format for multi-valued network ) formatwith the help of VIS / Vl2mv tool of Berkeley. In thispaper, we optimize on a number of standard industrialbenchmark circuit by MVSIS and ABC tool. Afteroptimization here used some technology mapping, thencompare the result. Here we try to find which tool giveoptimal result of optimization.

Index Terms- And-Inverter graph (AIG), DirectAcyclic Graph (DAG), Hardware Description Language(HDL), VHSIC Hardware Description Language(VHDL)

I. INTRODUCTIONOptimization of binary or multi-valued logic

networks using logic synthesis play an important role ina digital network system [1].

Figure-1: Design FlowLogic synthesis is used on a network which is

derived by compiling HDLs, VHDL or Verilog. Thentechnology mapping performed for standard cell orprogrammable devices. Logic synthesis is also fruitfulfor hardware emulation, design complexity estimation,software synthesis, and fast preprocessing of the circuitsbefore equivalence checking [2]. Figure-1 shows theprocess to get optimize result by MVSIS and ABC tool.

II. MVSISMVSIS is a sequel program modeled of SIS.

MVSIS is technology independent transformation ofcombinational as well as sequential logic system. Itworks on is such that all variables can be multi-valued,each with its own range [7]. MVSIS input formats canbe

1. PLA or BLIF : For Binary functions and networks2. BLIF-MV: For Multi-valued functions and networks

And for FSMs and finite automata three options areavailable

1. Using BLIF/BLIF-MV followed by “stg_extract”2. Using modified KISS2 format3. Using modified BLIF-MV format

To analyze the performance of this tool, scriptis applied over 15 combinational MCNC benchmarkcircuit. In MVSIS script logic synthesis is a sequence ofapplying optimization steps i.e. SWEEP (For removingredundant nodes), ELIMINATE and RESUBSTITUTE(For finding better logic boundaries), FAST_EXTRACT(For discovering shared logic boundary) and SIMPLIFYand FULL_SIMPLIFY (For simplifying the noderepresentation).[3]

In the Table-1, the first column shows the 15standard combinational MCNC benchmark circuit [4].Next seven columns show the statics of the benchmarkcircuit before applying the standard script. Here, PI isthe number of primary inputs; PO is the number ofprimary output, while Lits indicate the literals. The nextsection of the table shows, the reduced number of node,level, cubes, literals and literal(ff) after run MVSIS(script) over the benchmark.

Network Toplogy Reading

(VHDL/ Verilog - BV/MV)

BLIF / BLIF_MV Conversion

(VIs / Vl2mv)

Technology Independent Optimization

(SIS / MVSIS / ABC)

Application of Mapping Algorithm

(SIS / MVSIS / ABC)

Result Analysis






[email protected]














Network Toplogy Reading

(VHDL/ Verilog - BV/MV)

BLIF / BLIF_MV Conversion

(VIs / Vl2mv)

Technology Independent Optimization

(SIS / MVSIS / ABC)

Application of Mapping Algorithm

(SIS / MVSIS / ABC)

Result Analysis






[email protected]















TABLE-1STATICS OF BENCHMARK CIRCUIT BEFORE AND AFTER APPLYING THE MVSIS script

Bench-MarkCircuit

Circuit Statics After Run MVSIS ( script)PI PO Node Level Cubes Lits Lits(ff) Node Level Cubes Lits Lits(ff)

Duke2 22 29 385 8 590 992 992 362 8 772 898 890Rd84 8 4 495 9 894 1442 1442 495 10 1115 1278 1249Misex2 25 18 123 6 134 246 246 83 5 156 189 186B12 15 9 769 9 1214 1860 1860 110 7 229 279 269Cordic 23 2 2079 12 3283 4969 4969 1500 13 3414 4163 3909Pdc 16 40 2326 17 4203 8259 7364 3528 14 7821 10046 9140Spla 16 46 2237 18 4128 7961 7034 3424 13 7393 9173 8578C432 36 7 355 29 373 567 567 173 19 341 442 390C1355 41 32 949 44 949 1467 1467 220 10 428 684 678C1908 33 25 798 55 798 1245 1245 274 18 566 795 771T481 16 1 1075 11 1555 2673 2673 783 11 1945 2314 2142B9 41 21 275 20 351 408 408 98 6 175 220 204Dalu 75 16 2609 59 3595 4541 4541 743 15 1507 1962 1805Des 256 245 2263 10 3957 8991 8991 3823 12 7758 9864 9453K2 45 45 399 4 1521 3176 3176 1368 10 2850 3409 3366

TABLE-2DEGREE OF REDUCTION IN NODE AND LITERAL

BenchMarkCircuit

Circuit Statics AfterMVSIS (Script)

Degree of reduction

Node Literal Node Literal Node LiteralDuke2 385 992 362 898 0.060 0.095Rd84 495 1442 495 1278 0 0.114Misex2 123 246 83 189 0.325 0.231B12 769 1860 110 279 0.857 0.850Cordic 2079 4969 1500 4163 0.278 0.162Pdc 2326 8259 3528 10046 (-)0.517 (-)0.216Spla 2237 7961 3424 9173 (-)0.531 (-)0.152C432 355 567 173 442 0.513 0.220C1355 949 1467 220 684 0.768 0.534C1908 798 1245 274 795 0.657 0.386T481 1075 2673 783 2314 0.272 0.134B9 275 408 98 220 0.644 0.461Dalu 2609 4541 743 1962 0.715 0.562Des 2263 8991 3823 9864 (-)0.689 (-)0.097K2 399 3176 1368 3409 (-)2.429 (-)0.073

Table-2 shows the degree of reduction inbenchmark circuit. Here B12 circuit have maximumdegree of reduction in node as well as literal is 0.857 and0.850 respectively. In some of the benchmarks, i.e. PDC,SPLA, DES and K2, number of node and literals areincreased, but area is reduce as shown in the Table-3.

Table-3 shows the degree of area reduction and delayreduction when benchmark circuits mapped withmcnc.genlib. And, also show area optimization withLUT (#K=5).


TABLE-3ANALYSIS OF AREA AND DELAY BEFORE AND AFTER OPTIMIZATION

Becnch-MarkCircuit

Mapping with mcnc.genlib Mapping with LUT #k=5BeforeMVSIS script

AfterMVSISscript

Degree ofAreaReduction

DelayReduction

Area Beforeoptimization

Area Afteroptimization

Degree ofreduction

Area Delay Area DelayDuke2 978 9.3 896 9.3 0.084 0.0 220 207 0.059Rd84 1437 10.7 1249 11.2 0.131 (-)0.5 275 241 0.124Misex2 205 6.2 186 6.6 0.093 (-)0.4 42 34 0.190B12 1501 10.4 270 8.4 0.820 2.0 299 84 0.719Cordic 6384 14.3 3918 15.5 0.386 (-)1.2 647 551 0.148Pdc 10209 17 9193 16.7 0.100 0.3 1824 1639 0.101Spla 9534 16 8599 15.9 0.098 0.1 1839 1583 0.139C432 380 19.9 394 21.8 (-)0.037 (-)1.9 80 51 0.363C1355 816 15.6 788 15.9 0.034 (-)0.3 68 68 0.0C1908 852 24.1 844 23 0.009 1.1 114 107 0.061T481 2524 13.3 2147 13.3 0.149 0.0 410 343 0.163B9 229 7.2 205 6.7 0.105 0.5 42 40 0.048Dalu 2828 24.6 1847 18 0.347 6.6 382 276 0.277Des 11362 11.6 9562 13.7 0.158 2.1 1491 1332 0.107K2 3646 11.8 3409 12.1 0.065 (-)0.3 722 696 0.036

III. AIG REWRITING (ABC)An And-Inverter graph (AIG) is a direct acyclic

graph (DAG). In which node has either 0 or 2 incomingedges. If a node will have no edge then it will primaryinput (PI) or if node has twoinput edges then it will two-input AND gate. An edge become a normal input orcomplement of input. Some node will primary output(PO). Another logic synthesis tool is ABC. In whichAnd-Inverter Graphs (AIGs) can be rewrite for givennetwork. Rewriting is a fast greedy algorithm[2]. This isused for minimizing the AIG size by iteratively selectingAIG sub-graphs rooted at a node and replacing themwith smaller pre-computed sub-graphs, while preservingthe functionality of the root node. rewriting algorithm isdeveloped with following features:

1. Using 4-feasible cuts instead of two-level sub-graphs.

2. Restricting rewriting to preserve the number of logiclevels.

3. Developing several variations of AIG rewriting to4. Selectively collapse and refactor larger sub-graphs.5. Balance AIG using algebraic tree height reduction

In the AIG rewriting, the nodes are visited in atopological order. For each 4-input cut of a node, all pre-computed sub-graphs of its NPN class are considered.Logic sharing between the new sub-graph and nodesalready in the network is determined. First old sub-graphis dereference and the number of nodes, whose referencecounts became 0, is returned [6]. These nodes will beremoved if the old sub-graph is replaced. Next, a newsub-graph is added while counting the number of newnode and the node whose reference count went from 0 toa positive value. These nodes will add. The difference ofthe counter is the gain in the number of nodes if thereplacement is done. The new node is de-referenced andthe old node is referenced to return the AIG to itsoriginal state.

After trying all available sub-graphs for the givennode, the one that leads to the largest improvement at anode is used. If there is no improvement and “Zero-Costreplacement” is enabled, a new sub-graph that does notincrease the number of nodes is used. Table-4 shows thereduction in node with every iteration. Table-5 showsthe reduction in area and delay when technology mappedby mcnc.genlib.

TABLE-4REWRITING PERFORMANCE OF MCNC BENCH MARK

BenchMarkCircuit

First Iteration (rwz) Second Iteration (rwz) Third Iteration (rwz)NodeRewritten

Gain % Gain NodeRewritten

Gain % Gain NodeRewritten

Gain %Gain

Duke2 274 57 10.59 221 30 6.24 203 19 4.21Rd84 224 71 9.97 213 30 4.68 201 22 3.60Misex2 45 12 10.17 36 4 3.77 36 1 0.98B12 632 246 25.13 401 114 15.55 342 67 10.82Cordic 1164 1785 64.37 310 125 12.65 242 69 8.00


Pdc 2397 208 4.16 2082 208 4.34 1930 116 2.53Spla 2195 165 3.46 1904 219 4.76 1767 109 2.49C432 99 31 14.62 65 14 7.73 48 9 5.39C1355 220 100 19.53 59 18 4.37 54 0 0.0C1908 125 31 7.54 74 8 2.11 70 2 0.54T481 521 514 37.11 262 121 13.89 193 55 7.33B9 36 28 23.14 25 7 7.53 19 2 2.33Dalu 714 571 32.91 311 53 4.55 249 14 1.26Des 2003 993 20.99 1128 245 6.55 842 47 1.35K2 1470 458 20.01 1055 104 5.68 921 38 2.20

TABLE-5ANALYSIS OF ABC WITH MCNC.GENLIB

BenchMarkCircuit

Before ABC script.scr After ABC script.scr Degree of reductionNode Net Area Delay Node Net Area Delay Node Net Area Delay

Duke2 392 959 966 9.2 353 830 835 9.3 0.099 0.135 0.136 (-)0.1Rd84 511 1449 1452 10.7 497 1237 1239 11.3 0.027 0.146 0.147 (-)1.6Misex2 83 195 195 6.2 80 180 180 5.8 0.036 0.077 0.077 0.4B12 654 1786 1786 10.4 387 1011 1011 10.2 0.408 0.434 0.434 0.2Cordic 1970 6464 6465 14.3 1139 2982 2991 14.7 0.422 0.539 0.537 (-)0.4Pdc 3825 9558 9592 17.2 3389 8665 8694 16.4 0.114 0.093 0.094 0.8Spla 3712 9221 9250 16.2 3269 8175 8201 15.9 0.119 0.113 0.113 0.3C432 185 404 405 21.9 152 338 348 20.5 0.178 0.163 0.141 1.4C1355 218 480 816 15.6 200 431 769 14.1 0.083 0.102 0.058 1.5C1908 296 677 902 24.5 288 648 863 24.3 0.027 0.043 0.043 0.2T481 920 2558 2566 13.3 753 1993 1999 13.6 0.182 0.221 0.221 (-)0.3B9 99 216 217 7.8 87 181 182 7.0 0.121 0.162 0.161 0.8Dalu 1089 2578 2808 24.6 848 1980 2080 20.8 0.221 0.232 0.259 3.6Des 3687 9423 9718 12.7 3243 8210 8518 12.2 0.120 0.129 0.123 0.5K2 1411 3536 3581 12.4 1236 2997 3053 12.1 0.124 0.152 0.147 0.3

IV. ANALYSIS OF OPTIMIZATIONThis section compares AIG rewriting in ABC with

logic synthesis in MVSIS on MCNC benchmark. Table-6 shows that Node reduction with MVSIS (script) andABC (resyn2), and also shows that area and delay

reduction of mapping of optimize benchmark, whichshows it is better in ABC. ABC (resyn2) based on areaoptimization under delay constraints. Therefore in somebenchmarks delay is increased as compare to MVSIS(script).

TABLE-6COMPARISON BETWEEN MVSIS AND ABC

Bench Mark Circuit Degree of Reduction in MVSIS Degree of Reduction in ABCNode Area Delay Node Area Delay

Duke2 0.060 0.084 0.0 0.099 0.136 (-)0.1Rd84 0 0.131 (-)0.5 0.027 0.147 (-)1.6Misex2 0.325 0.093 (-)0.4 0.036 0.077 0.4B12 0.857 0.820 2.0 0.408 0.434 0.2Cordic 0.278 0.386 (-)1.2 0.422 0.537 (-)0.4Pdc (-)0.517 0.100 0.3 0.114 0.094 0.8Spla (-)0.531 0.098 0.1 0.119 0.113 0.3C432 0.513 (-)0.037 (-)1.9 0.178 0.141 1.4C1355 0.768 0.034 (-)0.3 0.083 0.058 1.5C1908 0.657 0.009 1.1 0.027 0.043 0.2T481 0.272 0.149 0.0 0.182 0.221 (-)0.3B9 0.644 0.105 0.5 0.121 0.161 0.8Dalu 0.715 0.347 6.6 0.221 0.259 3.6Des (-)0.689 0.158 2.1 0.120 0.123 0.5K2 (-)2.429 0.065 (-)0.3 0.124 0.147 0.3

V. CONCLUSIONAIG rewriting is an innovative technique for

combinational logic synthesis. This experiment shows

that AIG rewriting often leads to quality comparableor better than those afforded by the logic synthesisscript in MVSIS. The extreme speed and good quality


of the proposed algorithm might make the new flowuseful in a variety of applications such as hardwareemulation, estimation of design complexity, andequivalence checking.

REFERENCE

[1]. ‘Optimization of Multi-Valued Multi-LevelNetworks’- M.Gao, J-H. Jiang, Y. Jiang, Y. Li, A.Mishchenko, S. Sinha, T. Villa and R. Brayton; 32nd

IEE International Symposium on Multiple-ValuedLogic (ISMVL’02

[2]. ‘DAG-Aware AIG Rewriting – A Fresh Look atcombinational Logic Synthesis’ – Alan Mishchenko,Satrajit Chatterjee, Robert Brayton; DAC 2006

[3]. ‘Minimization of Multiple Valued Functions in PostAlgebra’ – Elena Dubrova, Yunjian Jiang, RobertBrayton

[4]. ‘Logic Synthesis and Optimization Benchmarks UserGuide- Version 3.0’ - Saeyang Yang

[5]. ‘Multi-Valued Logic Synthesis’ Robert K Brayton,Sunil P Khatri[6]. ‘Quick Look under the Hood of ABC – AProgrammer’s Manual’[7]. http://www.eecs.berkeley.edu

International Journal of Electrical & Electronics · PDF fileInternational Journal of...

Documents

Transcript of International Journal of Electrical & Electronics · PDF fileInternational Journal of...