IJITCE Oct 2011

UK: Managing Editor International Journal of Innovative Technology and Creative Engineering 1a park lane, Cranford London TW59WA UK E-Mail: [email protected] Phone: +44-773-043-0249

USA: Editor International Journal of Innovative Technology and Creative Engineering Dr. Arumugam Department of Chemistry University of Georgia GA-30602, USA. Phone: 001-706-206-0812 Fax:001-706-542-2626

India: Editor International Journal of Innovative Technology & Creative Engineering Dr. Arthanariee. A. M Finance Tracking Center India 261 Mel quarters Labor colony, Guindy, Chennai -600032. Mobile: 91-7598208700

www.ijitce.co.uk

IJITCE PUBLICATION

INTERNATIONAL JOURNAL OF INNOVATIVE

TECHNOLOGY & CREATIVE ENGINEERING

Vol.1 No.10

October 2011

www.ijitce.co.uk

From Editor's Desk

Dear Researcher, Greetings! This monthly journal contains new research topics in the field of RHA under elevated temperature, visual tracking system with UAV, Heat transfer on MHD peristaltic transport. We would like to dedicate this month journal to the visionary co-founder of Apple who helped usher in the era of personal computers and then led a cultural transformation in the way music, movies and mobile communications were experienced in the digital age. Let us review world research focus for this month. Aircraft designers test their creations in a wind tunnel and now spacecraft designers can do the same in a "space tunnel". Researchers at the German Aerospace Center (DLR) in Göttingen have built a 12-metre-long vacuum chamber designed to replicate the conditions of outer space at a few degrees above absolute zero. The chamber is large enough to test entire sections of satellites and will be used to research ion engines and other electric spacecraft propulsion systems. The ion beams from these engines can hit sensitive electrical parts of satellites, such as solar panels, and cause them to fail. The researchers hope that a better understanding of the beams will reduce the damage they cause. There is new hope for heavy smokers, people with asthma and those with chronic lung scarring. Stem cells have been discovered that rapidly rebuild alveoli, the tiny air sacs in lungs. The key is that the blood vessels turn on the pathways for regeneration, says Shahin Rafii of Weill Cornell Medical College in New York City, who led the team. Aspirin dramatically reduces the risk of developing colorectal cancer in people with a family history of the disease, providing the most direct evidence yet that the drug can be used for cancer prevention. While previous studies have hinted that aspirin might prevent cancer, this is the first study where the primary goal was to look at whether the drug reduced the risk of cancer. "We set out to see if aspirin would prevent cancer, and it does," says John Burn of Newcastle University, UK, who led the study. Bangkok's main river broke its banks overnight forcing thousands of residents to flee the flooding Thai capital. It is Thailand's worst flooding in fifty years. Since mid July 373 people have died, and the waters have disrupted the lives of nearly 2.5 million. More than 113,000 people are in shelters and 720,000 people seeking medical attention. How are we going to handle these kinds of disasters in future? Do we have backup plan for such situations? It puts people in a new kind of environment, which they never experienced before. How are we going to train people minds to handle these kinds of situations in other parts of the world? As a researcher, how do you think you can handle this? We can look at the impact of every product, which undergoes this situation, for example, starting from your car. What are the changes we need to make to our every day products so that it can be used during disaster situations? It has been an absolute pleasure to present you articles that you wish to read. We look forward to many more new technology-related research articles from you and your friends. We are anxiously awaiting the rich and thorough research papers that have been prepared by our authors for the next issue. Thanks, Editorial Team IJITCE

Editorial Members Dr. Chee Kyun Ng Ph.D Department of Computer and Communication Systems, Faculty of Engineering, Universiti Putra Malaysia,UPM Serdang, 43400 Selangor,Malaysia. Dr. Simon SEE Ph.D Chief Technologist and Technical Director at Oracle Corporation, Associate Professor (Adjunct) at Nanyang Technological University Professor (Adjunct) at Shangai Jiaotong University, 27 West Coast Rise #08-12,Singapore 127470 Dr. sc.agr. Horst Juergen SCHWARTZ Ph.D, Humboldt-University of Berlin, Faculty of Agriculture and Horticulture, Asternplatz 2a, D-12203 Berlin, Germany Dr. Marco L. Bianchini Ph.D Italian National Research Council; IBAF-CNR, Via Salaria km 29.300, 00015 Monterotondo Scalo (RM), Italy Dr. Nijad Kabbara Ph.D Marine Research Centre / Remote Sensing Centre/ National Council for Scientific Research, P. O. Box: 189 Jounieh, Lebanon Dr. Aaron Solomon Ph.D Department of Computer Science, National Chi Nan University, No. 303, University Road, Puli Town, Nantou County 54561, Taiwan Dr. Arthanariee. A. M M.Sc.,M.Phil.,M.S.,Ph.D Director - Bharathidasan School of Computer Applications, Ellispettai, Erode, Tamil Nadu,India Dr. Takaharu KAMEOKA, Ph.D Professor, Laboratory of Food, Environmental & Cultural Informatics Division of Sustainable Resource Sciences, Graduate School of Bioresources, Mie University, 1577 Kurimamachiya-cho, Tsu, Mie, 514-8507, Japan Mr. M. Sivakumar M.C.A.,ITIL.,PRINCE2.,ISTQB.,OCP.,ICP Project Manager - Software, Applied Materials, 1a park lane, cranford, UK Dr. Bulent Acma Ph.D Anadolu University, Department of Economics, Unit of Southeastern Anatolia Project(GAP), 26470 Eskisehir, TURKEY Dr. Selvanathan Arumugam Ph.D Research Scientist, Department of Chemistry, University of Georgia, GA-30602, USA.

Review Board Members

Dr. T. Christopher, Ph.D., Assistant Professor & Head,Department of Computer Science,Government Arts College(Autonomous),Udumalpet, India. Dr. T. DEVI Ph.D. Engg. (Warwick, UK), Head,Department of Computer Applications,Bharathiar University,Coimbatore-641 046, India. Dr. Giuseppe Baldacchini ENEA - Frascati Research Center, Via Enrico Fermi 45 - P.O. Box 65,00044 Frascati, Roma, ITALY. Dr. Renato J. orsato Professor at FGV-EAESP,Getulio Vargas Foundation,São Paulo Business School,Rua Itapeva, 474 (8° andar) ,01332-000, São Paulo (SP), Brazil Visiting Scholar at INSEAD,INSEAD Social Innovation Centre,Boulevard de Constance,77305 Fontainebleau - France Y. Benal Yurtlu Assist. Prof. Ondokuz Mayis University Dr. Paul Koltun Senior Research ScientistLCA and Industrial Ecology Group,Metallic & Ceramic Materials,CSIRO Process Science & Engineering Private Bag 33, Clayton South MDC 3169,Gate 5 Normanby Rd., Clayton Vic. 3168 Dr.Sumeer Gul Assistant Professor,Department of Library and Information Science,University of Kashmir,India Chutima Boonthum-Denecke, Ph.D Department of Computer Science,Science & Technology Bldg., Rm 120,Hampton University,Hampton, VA 23688

Dr. Renato J. Orsato Professor at FGV-EAESP,Getulio Vargas Foundation,São Paulo Business SchoolRua Itapeva, 474 (8° andar), 01332-000, São Paulo (SP), Brazil Lucy M. Brown, Ph.D. Texas State University,601 University Drive,School of Journalism and Mass Communication,OM330B,San Marcos, TX 78666 Javad Robati Crop Production Departement,University of Maragheh,Golshahr,Maragheh,Iran Vinesh Sukumar (PhD, MBA) Product Engineering Segment Manager, Imaging Products, Aptina Imaging Inc. doc. Ing. Rostislav Choteborský, Ph.D. Katedra materiálu a strojírenské technologie Technická fakulta,Ceská zemedelská univerzita v Praze,Kamýcká 129, Praha 6, 165 21 Dr. Binod Kumar M.sc,M.C.A.,M.Phil.,ph.d, HOD & Associate Professor, Lakshmi Narayan College of Tech.(LNCT), Kolua, Bhopal (MP) , India. Dr. Paul Koltun Senior Research ScientistLCA and Industrial Ecology Group,Metallic & Ceramic Materials,CSIRO Process Science & Engineering Private Bag 33, Clayton South MDC 3169,Gate 5 Normanby Rd., Clayton Vic. 3168 DR.Chutima Boonthum-Denecke, Ph.D Department of Computer Science,Science & Technology Bldg.,Hampton University,Hampton, VA 23688 Mr. Abhishek Taneja B.sc(Electronics),M.B.E,M.C.A., M.Phil., Assistant Professor in the Department of Computer Science & Applications, at Dronacharya Institute of Management and Technology, Kurukshetra. (India). doc. Ing. Rostislav Chot ěborský,ph.d, Katedra materiálu a strojírenské technologie, Technická fakulta,Česká zemědělská univerzita v Praze,Kamýcká 129, Praha 6, 165 21 Dr. Amala VijayaSelvi Rajan, B.sc,Ph.d, Faculty – Information Technology Dubai Women’s College – Higher Colleges of Technology,P.O. Box – 16062, Dubai, UAE Naik Nitin Ashokrao B.sc,M.Sc Lecturer in Yeshwant Mahavidyalaya Nanded University Dr.A.Kathirvell, B.E, M.E, Ph.D,MISTE, MIACSIT, MEN GG Professor - Department of Computer Science and Engineering,Tagore Engineering College, Chennai Dr. H. S. Fadewar B.sc,M.sc,M.Phil.,ph.d,PGDBM,B.Ed . Associate Professor - Sinhgad Institute of Management & Computer Application, Mumbai-Banglore Westernly Express Way Narhe, Pune - 41 Dr. David Batten Leader, Algal Pre-Feasibility Study,Transport Technologies and Sustainable Fuels,CSIRO Energy Transformed Flagship Private Bag 1,Aspendale, Vic. 3195,AUSTRALIA Dr R C Panda (MTech & PhD(IITM);Ex-Faculty (Curtin Univ Tech, Perth, Australia))Scientist CLRI (CSIR), Adyar, Chennai - 600 020,India Miss Jing He PH.D. Candidate of Georgia State University,1450 Willow Lake Dr. NE,Atlanta, GA, 30329 Dr. Wael M. G. Ibrahim Department Head-Electronics Engineering Technology Dept.School of Engineering Technology ECPI College of Technology 5501 Greenwich Road - Suite 100,Virginia Beach, VA 23462 Dr. Messaoud Jake Bahoura Associate Professor-Engineering Department and Center for Materials Research Norfolk State University,700 Park avenue,Norfolk, VA 23504 Dr. V. P. Eswaramurthy M.C.A., M.Phil., Ph.D., Assistant Professor of Computer Science, Government Arts College(Autonomous), Salem-636 007, India. Dr. P. Kamakkannan,M.C.A., Ph.D ., Assistant Professor of Computer Science, Government Arts College(Autonomous), Salem-636 007, India. Dr. V. Karthikeyani Ph.D., Assistant Professor of Computer Science, Government Arts College(Autonomous), Salem-636 008, India. Dr. K. Thangadurai Ph.D., Assistant Professor, Department of Computer Science, Government Arts College ( Autonomous ), Karur - 639 005,India.

Dr. N. Maheswari Ph.D., Assistant Professor, Department of MCA, Faculty of Engineering and Technology, SRM University, Kattangulathur, Kanchipiram Dt - 603 203, India. Mr. Md. Musfique Anwar B.Sc(Engg.) Lecturer, Computer Science & Engineering Department, Jahangirnagar University, Savar, Dhaka, Bangladesh. Mrs. Smitha Ramachandran M.Sc(CS)., SAP Analyst, Akzonobel, Slough, United Kingdom. Dr. V. Vallimayil Ph.D., Director, Department of MCA, Vivekanandha Business School For Women, Elayampalayam, Tiruchengode - 637 205, India. Mr. M. Rajasenathipathi M.C.A., M.Phil Assistant professor, Department of Computer Science, Nallamuthu Gounder Mahalingam College, India. Mr. M. Moorthi M.C.A., M.Phil., Assistant Professor, Department of computer Applications, Kongu Arts and Science College, India Prema Selvaraj Bsc,M.C.A,M.Phil Assistant Professor,Department of Computer Science,KSR College of Arts and Science, Tiruchengode Mr. V. Prabakaran M.C.A., M.Phil Head of the Department, Department of Computer Science, Adharsh Vidhyalaya Arts And Science College For Women, India. Mrs. S. Niraimathi. M.C.A., M.Phil Lecturer, Department of Computer Science, Nallamuthu Gounder Mahalingam College, Pollachi, India. Mr. G. Rajendran M.C.A., M.Phil., N.E.T., PGDBM., P GDBF., Assistant Professor, Department of Computer Science, Government Arts College, Salem, India. Mr. R. Vijayamadheswaran, M.C.A.,M.Phil Lecturer, K.S.R College of Ars & Science, India. Ms.S.Sasikala,M.Sc.,M.Phil.,M.C.A.,PGDPM & IR., Assistant Professor,Department of Computer Science,KSR College of Arts & Science,Tiruchengode - 637215 Mr. V. Pradeep B.E., M.Tech Asst. Professor, Department of Computer Science and Engineering, Tejaa Shakthi Institute of Technology for Women, Coimbatore, India. Dr. Pradeep H Pendse B.E.,M.M.S.,Ph.d Dean - IT,Welingkar Institute of Management Development and Research, Mumbai, India Mr. K. Saravanakumar M.C.A.,M.Phil., M.B.A, M.Tech, PGDBA, PGDPM & IR Asst. Professor, PG Department of Computer Applications, Alliance Business Academy, Bangalore, India. Muhammad Javed Centre for Next Generation Localisation, School of Computing, Dublin City University, Dublin 9, Ireland Dr. G. GOBI Assistant Professor-Department of Physics,Government Arts College,Salem - 636 007 Dr.S.Senthilkumar Research Fellow,Department of Mathematics,National Institute of Technology (REC),Tiruchirappli-620 015, Tamilnadu, India.

Contents 1. Phylogenetic Characterization Of Human Genome Through DNA Fingerprinting

Among Indians - Adlin Blessi.T.L., ……….[1]

2. Location Privacy Using User Anonymity And Dummy Locations - Kusum Gupta, Ajay Singh Yadav and Shashank Yadav ……………..[5]

3. Performance of Mortar Incorporating RHA Under Elevated Temperature - Muhammad Harunur Rashid, Md. Keramat Ali Molla, Tarif Uddin Ahmed ……….………………………..…[9]

4. Implementation of an Onboard Visual Tracking System with Small Unmanned Aerial Vehicle (UAV) - Ashraf Qadir, William Semke, Jeremiah Neubert …………...[17]

INTERNATIONAL JOURNAL OF INNOVATIVE TECHNOLOGY & CREATIVE ENGINEERING (ISSN:2045-8711) VOL.1 NO.10 OCTOBER 2011

1

Phylogenetic Characterization Of Human Genome Through DNA Fingerprinting Among

Indians Adlin Blessi.T.L.,

Department of Biotechnology, Karunya University, Coimbatore, India. [email protected]

Abstract—DNA fingerprinting technique is considered as a powerful technique and is widely used all ov er the globe today. The basic requirement for it is the availability of biological samples of the individua ls. Buccal cells are increasingly used as a source of q uality DNA to improve participation rates in molecular stu dies. These cells are routinely shed and replaced by new cells. We used Random Amplified Polymorphic DNA (RAPD) markers to assess genetic diversity among 6 individ uals. The amount of genetic variation was evaluated by polymerase chain reaction amplification with 10-mer oligonucleotide primers. The similarity co-efficient between each pair of accession were used to constru ct a Dendrogram showing relationships between them using Unweighed Pair Group Method Average (UPGMA).

Keywords: Buccal cells, DNA Fingerprinting, Random Amplified Polymorphism (RAPD), and genetic variation.

I. INTRODUCTION

The characterization of a DNA sample for individual identity according to its chemistry or sequence information often referred to as DNA "fingerprinting” [1]. The preferred way to obtain genomic DNA is from peripheral blood [2]. Blood sampling, however, may be problematic in cases such as extreme illness or elderly people, babies and people who are unwilling to this invasive procedure [3]. In contrast to blood biological samples such as buccal cells can supply DNA for genetic testing and provide a noninvasive approach [4]. Several cell collection methods are there but mouthwash procedure gives high yield of DNA from buccal cell [5]. DNA Fingerprinting can be defined as a technique that is used for revealing the identity of an organism at the molecular level. Usually fingerprinting is based on the morphological features and is restricted to humans but DNA fingerprinting is a technique of finding the genetic diversity. This is primarily based on the polymorphisms occurring at the molecular level that is on the base sequences of the genome [6]. RAPD based molecular characterization is an important tool to explore genetic biodiversity between morphologically identical species and genetic relatedness between distant species [7]. The random amplified polymorphic DNA (RAPD) method is based on the polymerase chain reaction (PCR) using short (usually 10 nucleotide)

primers of arbitrary sequences [8]. Since the technique is relatively easy to apply to a wide array of plant and animal taxa, and the number of loci that can be examined is essentially unlimited, RAPDs are viewed as having several advantages over RFLPs and DNA fingerprints [9]. The RAPD technique has several advantages such as the ease and rapidity of analysis, a relatively low cost, availability of a large number of primers and the requirement of a very small amount of DNA for analysis [10]. DNA purity plays an important role in amplification patterns obtained by RAPD-PCR [11]. Since the primers consist of random sequences, and do not discriminate between coding and noncoding regions, it is reasonable to expect the technique to sample the genome more randomly than conventional methods [9].

In this study DNA is isolated from buccal cells of individuals from different ethnic background. Six subjects from different ethnic background were chosen and the genetic variation among them is analysed by using dendrogram analysis.

II. MATERIALS AND METHODS

A. Collection of buccal cells:

Buccal cells were collected fresh before DNA isolation. One hour after brushing the subjects were asked to scrape the inner cheeks & rinse the mouth with a 10 ml of sterile distilled water for 60seconds and asked to collect the mouthwash in a beaker and transfer to 15 ml centrifuge tubes. During this period the subjects were instructed not to eat anything.

B. DNA Isolation

To the mouthwash collected, 3ml of TNE solution [17 mM Tris/HCl (pH 8.0), 50 mM NaCl and 7 mM EDTA] diluted in 66% ethanol was added and centrifuged for 15 min at 7000 rpm at room temperature & the supernatant was discarded immediately. For second washing 1 ml of TNE was added to resuspend the cells and centrifuged at 7000 rpm for 10 min and the supernatant was discarded. The cell pellet was vortexed vigorously for 5 seconds and 1.3 ml of lysis solution [10mM Tris (pH8.0), 0.5% SDS, 5mM EDTA] and 15 µL of proteinase K were


2

added. The mixture was vortexed for 5 seconds at medium speed, followed by incubation for 4hrs at 55°C then 1.4 ml of the mixture was transferred to a 2 ml micro-centrifuge tube. 600µl of a solution containing 8 M ammonium acetate and1 mM EDTA was added, to remove Proteins and other contaminants and vortexing was done at high speed for 5 seconds and centrifuged at 10,000 rpm for 10 min. 900µl of supernatant was carefully poured into two clean 2ml micro-centrifugetubes containing 540µl of isopropanol and is kept for overnight refrigeration. The tubes were gently inverted 20 times & centrifuged at 10000rpm for 10 min, the supernatant

was discarded and the tubes were drained. 2ml of 70% ice cold ethanol was added and mixed by inverting and centrifuged at 10000rpm for 10 min. & the supernatant was discarded. The tube was inverted and drained on clean absorbent paper, then allowed to air dry for 45 to 60 min. The DNA was re-suspended in 10-40µl of TE buffer [10mMTris (pH7.8) and 1mM EDTA]. The DNA sample was run in 0.8% agarose gel and the bands was visualized under UV light using UV Transilluminator and it is then photographed in Gel documentation.

C. RAPD Analysis

The Polymerised Chain Reaction starts with denaturation of the double stranded DNA to form single strands (initial denaturation- 94°C for 5 min and denaturation-94°C for 40 sec) followed by annealing (36°C for 30 sec) and extension steps (extension-72 °C for 90 sec & final extension-72°C for 10 min). Amplification was carried out with a 50µL reaction mixture containing Primer (2µM/µL) -8.0µL, 10X Buffer-5.0 µL, 2mM dNTP Mix- 5.0µL, Taq DNA polymerase (5U/µL)- 0.5µL, Template DNA (50ng)- 2.0µL, Sterile distilled water- 29.5µL. The primers used were PG05 – 5’ GCAGGCTAAC 3’ PG06 – 5’ CCTGGTGGTC 3’ PG07 – 5’ GCTGCAGTAG 3’. About 34 cycles of reactions were carried out and was run in a 2% agarose gel electrophoresis and visualized using UV transilluminator and is photographed.

D. Creating Dendrogram

The data obtained from amplification products by primers were used to estimate genetic similarity among different isolates on the basis of shared amplification products. The RAPD patterns were scored on the basis of presence or absence of band. The similarity coefficients were utilized to generate Dendrogram by

using UPGMA (Unweighed Pair Group Method of Arithmetic means) through the programme, Phylip Version 3.69.

III. RESULTS AND DISCUSSION

The isolates of Homo sapiens were isolated from different states of India. The samples were collected from people belonging to Jharkhand, Orissa, Kerala, Andhra Pradesh, and Tamil Nadu. DNA Finger printing was done and the RAPD banding pattern was observed stable for each population of the primer. Out of PG 05, PG 06, PG 07, the primer PG 06 alone gave the band variation. The primer PG 05 and PG 07 gave only few bands. Thus the phylogenetic characterization was done through DNA Fingerprinting.

Several protocols have been developed to obtain DNA from buccal cells, but cell collection by mouthwash seems to give higher yields than many other methods. In a similar study conducted, it is revealed that the buccal cells provided usable amounts of DNA, particularly from mouthwash and cytobrush. Previous studies reported by Garcia-Closas et al,

[12] shows that the greatest efficiency for DNA extraction was from mouthwash.

The gels scored for computer analysis on the basis of the presence or absence of the amplified products. If the product was present in a genotype, it was designed as ‘0’. A total of 46 bands were observed from six individuals screened from various ethnic backgrounds. The similarity co-efficient between each pair of accession were used to construct a Dendrogram using Unweighed Pair Group Method Average (UPGMA). Cluster analysis was performed on the similarity index calculated by RAPD markers. Through UPGMA programme, the resultant Dendrogram was presented in Figure 3. Figure 3 depicts the analysis of banding patterns & number of polymorphic fragments ranged from 603 to 6557 bp. The polymorphism observed was 39.13 %.

The upper cluster consists of sample of Homo sapiens which was isolated from people belonging to Jharkhand (BC1). Lower cluster consists of samples which were isolated from people belonging Tamil Nadu (BC5), Andhra Pradesh (BC4, BC6), Kerala (BC3), Orissa (BC2). Sample collected from Andhra is closely related to Tamil Nadu and Orissa and only distantly related to sample collected from Jharkhand (BC1). Thus the geographical analysis was done from North to the southern regions of India.

In the present study the RAPD markers were more striking because a large number of markers were generated per primer at DNA level for 6 representative isolates. This technique was relatively simple and appeared more effective and positive since genetic diversity can be identified in terms of number of bands based on their presence or absence. The number of


3

band interpretation is given in Fig. 3., by comparing each samples, it is found that more number of bands were viewed in sample 3. This was in accordance with the findings of Smith et al. They reported that RAPD provide useful tools for diagnostic studies in the population, strains and species level. Different thermal cyclers [14], temperature profiles, band of DNA polymerases [15] and the concentration of MgCl2, primer and template DNA can affect the reproducibility of RAPD assay [13]. In our work, we standardised all the above parameters prior to performing our analysis. Stephen J et al., [16] described that animals from near ecological zone show more similarities which is similar to the results of our study represented in Dendrogram (Fig. 3.).

Studies carried out by GU et al. [17] revealed that single canine DNA bands derived from the RAPD assay may contain more than one DNA fragment of similar size, but a different nucleotide sequence. Consistent with this idea, it is possible that each of the amplified fragments could represent more than one DNA sequence. As described by Chansiripornchai et al. [18]. The genetic similarity matrix of RAPD data for 5 isolates was constructed based on coefficient of similarity utilized for the construction of Dendrogram using the UPGMA.

The study revealed that the GC rich primer PG 06 generated maximum number of products. Sample BC1 (Jharkhand) and BC3 (Kerala) showed vast genetic diversity, this may be due to climatic factors, food habits etc. Samples BC2 (Orissa), BC6 (Andhra Pradesh) and BC5 (Tamil Nadu), BC4 (Andhra Pradesh) are closely related to each other. This study revealed the phylogenetic reconstruction is an important approach in understanding the genetic diversity.

A. Figures

Fig. 1. Agarose gel (0.8%) electrophoresis showing total genomic DNA of Human obtained from mouthwashes Marker (M) – DNA Marker (1 kb Ladder) Lane 1 – Jharkhand, Lane 2 – Orissa, Lane 3 – Kerala Lane 4 – Andhra Pradesh Lane 5 – Tamil Nadu Lane 6 – Andhra Pradesh

Fig. 2. Agarose gel (2%) electrophoresis showing RAPD pattern of human genomic DNA Marker (M) – DNA Marker (λ DNA-HindIII and Φ X 174 DNA Hae III digest Mix) Lane 1 – Jharkhand Lane 2 – Orissa


4

Lane 3 – Kerala Lane 4 – Andhra Pradesh Lane 5 – Tamil Nadu Lane 6 – Andhra Pradesh

Fig. 3. Dendrogram depicting the genetic diversity among six DNA samples

IV. CONCLUSION

This study shows that genetic diversity among people occurs due to climatic factors and food habits too. The variation among different individuals can be studied through RAPD analysis and the phylogenetic characterization can be arrived.

AKNOWLEDGEMENT

First and foremost, I praise and thank Almighty God whose blessings have bestowed in me, the will power and confidence to carry out my project. I feel it a pleasure to be indebted to my guide, Dr.Philomena George (Professor), Department of Biotechnology for Her invaluable support, advice and encouragement. And I would also like to thank my parents and friends for their support.

REFERENCES [1] Gustavo Caetano-Anollés, Brant J. Bassam, and Peter M.

Gresshoff, “DNA Amplification Fingerprinting: A Strategy for Genome Analysis,” Plant Molecular Biology Reporter, vol.9(4), pp. 294-307, 1991.

[2] Heather Spencer Feigelson, Carmen Rodriguez, Andrea S. Robertson, Eric J. Jacobs, Eugenia E. Calle, Yvonne A. Reid, and Michael J. Thun, “Determinants of DNA Yield and Quality from Buccal Cell Samples Collected with Mouthwash,” Cancer Epidemiology, Biomarkers & Prevention, vol. 10, pp.1005–1008, 2001.

[3] Marisi Aidar, Sergio Roberto Peres Line, “A Simple and Cost-Effective Protocol for DNA Isolation from Buccal Epithelial Cells”, Braz. Dent.J., vol. 18(2), pp. 148-152, 2007.

[4] Claire Mulot, Isabelle St¨ucker, Jacqueline Clavel, Philippe Beaune, and Marie-Anne Loriot., “Collection of Human Genomic DNA from Buccal Cells for Genetics Studies: Comparison Between Cytobrush, Mouthwash, and Treated

Card,” Journal of Biomedicine and Biotechnology, Vol.3, pp.291–296, 2005.

[5] Cozier YC, Palmer JR, Rosenberg L., “Comparison of methods for collection of DNA samples by mail in the Black Women’s Health Study,” Ann Epidemiol., vol.2, pp.117-122, 2004.

[6] Kahl,G., “The Dictionary of Gene Technology,” Wiley-VCH, Weinheim, 2001.

[7] Deepika Abrol and S. K. Bhatnagar, “Biodiversity of few Indian charophyte taxa based on molecular characterization and construction of phylogenetic tree,” African Journal of Biotechnology, vol. 5(17), pp. 1511-1518, 2006.

[8] Lisa Monna, Akio Miyao, Takakazu Inoue, Shuichi Fukuoka, Muneo Yamazaki, Hui Sun Zhong, Takuji Sasaki, and Yuzo Minobe, “Determination of RAPD Markers in Rice and their Conversion into Sequence Tagged Sites (STSs) and STS-Specific Primers”, DNA Research, vol. 1, pp.139-148, 1994.

[9] Lynch.M and B.G.Milligan, “Analysis of population genetic structure with RAPD markers,” Molecular Ecology, vol.3, pp.91-99, 1994.

[10] Niknejad.A, M. A. Kadir, S. B. Kadzimin, N. A. P. Abdullah and K. Sorkheh, “Molecular characterization and Phylogenetic relationships among and within species of Phalaenopsis (Epidendroideae: Orchidaceae) based on RAPD analysis,” African Journal of Biotechnology, vol.8 (20), pp.5225-5240, 2009.

[11] Ching Wen Tseng, Rodney J. Sobieski and Scott S. Crupper, “Genotyping by Randomly Amplified Polymorphic DNA Is Influenced by the Method of DNA Preparation,” Transactions of the Kansas Academy of Science, vol. 104(1/2), pp.37-43, 2001.

[12] Garcia-Closas, M., K.M. Egan, J. Abruzzo, P.A. Newcomb, L. Titus-Ernstoff, T. Franklin, P.K. Bender, J.C. Beck, et al., “Collection of genomic DNA from adults in epidemiological studies by buccal cytobrush and mouthwash.” Cancer Epidemiol. Biomark. Prev., vol.10, pp. 687-696, 2001.

[13] Muralidharan K, Wakeland EK., "Concentration of primer and template qualitatively affects products in random amplified polymorphic DNA PCR,” Biotechniques, vol.14, pp.362-364, 1993.

[14] Edwards KJ., “Randomly amplified polymorphic DNAs (RAPDs),” Molecular Tools for Screening Biodiversity-Plants and Animals, pp.171-179,1998.

[15] Schierwater B, Ender A., “Different thermostable DNA polymerases may amplify different RAPD products,” Nucl. Acids Res., vol.21, pp.4647-4648, 1993.

[16] Stephens JC, Gilbert DA, Yuhki N, O’Brien SJ., “Estimation of heterozygosity for single-probe multilocus DNA fingerprints,” Molecular biology and Evolution, vol.9, pp.729-743, 1992.

[17] Gu W., Post M., Agurrie G.D., Ray K., “Individual bands obtained by RAPD analysis of canine genomic DNA often contain multiple DNA sequences,” The Am. Genet. Assoc., vol. 90, pp.96-98, 1999.

[18] Chansiripornchai, N., P. Ramasoota, A. Bangtrakuknonth, J. Sasipreeyajan and S.B.Svenson, “Application of randomly amplified polymorphic DNA (RAPD) analysis for typing avian Salmonella enterica subsp. Enteric,” FEMS Immunol. Med. Microbiol., Vol. 29, pp.221-225, 2000.


5

Location Privacy Using User Anonymity And Dummy Locations

Kusum Gupta, Ajay Singh Yadav and Shashank Yadav ([email protected],[email protected],[email protected])

Abstract-This paper concentrates on location privacy, a particular type of information privacy t hat can be defined as the ability to prevent others fro m learning one’s current or past location. Here we ar e proposing a new technique that uses user anonymity and dummy locations for location privacy while using location aware application server. User communicates with the server through a trusted proxy server. It sends dummy locations to the application server with its original position. The user uses temporary pseudonyms that are changed frequently according to some algorithm. Whenever pseudonyms are changed by a user, dummy locations are chosen in a tricky fashion. That make s the task of tracing the user very difficult

Keywords- Location privacy, pseudonym, de-anonymize, dummy-Locations, location anonymity, trusted proxy.

I. INTRODUCTION The Indian Constitution of 1950 does not expressly recognize the right to privacy. However, the Supreme Court first recognized it in 1964 that there is a right of privacy implicit in the Constitution under Article 21 of the Constitution, which states, "No person shall be deprived of his life or personal liberty except according to procedure established by law." The 1948 Universal Declaration of Human Rights [1] declares that everyone has a right to privacy at home, with family, and in correspondence.

This paper concentrates on location privacy. Location privacy is a type of information privacy that can be defined as the ability to prevent others from learning one’s current or past locations [1]. Location privacy is more important in pervasive computing environment. That implicitly implies that the communication device is mobile and wireless. User might not care if someone finds out where she was yesterday at 10:30 a.m., but if this someone could inspect the history of all her movements, recorded every second with great accuracy, might prove dangerous.

II. LOCATION DEPENDENT (AWARE) APPLICATIONS

To protect the privacy of our location information while taking advantage of location-aware services, we wish to hide our true identity from the applications receiving our location; at a very high level, this can be taken as a statement of our security policy. Now we try to develop a more sophisticated system for location-based service (Fig1) [1]. Here the user accesses the application server through a trusted proxy server. The user is authorized to use the service by this trusted proxy. The proxy and the user decide a pseudonym for the user. User sends the location and the requested service to proxy. The trusted proxy sends location and the request to the application server with user’s pseudonym. Response from the application server reaches the user through proxy. There are many users requesting the service through proxy. Proxy has to maintain a table for the user ID and the corresponding pseudonym redirect the response from the application to appropriate user. Here the application is aware of the location and the request from the user but doesn’t know her identity. Pseudonyms are changed frequently. So indirectly location privacy is gained.

Fig1. Location privacy using Temporary pseudonyms and a trusted proxy.


6

III. AN IMPROVEMENT USING DUMMY- LOCATIONS (DL).

Problem with previously discussed solution is that, if the system’s spatial and temporal resolution were sufficiently high, applications could easily link the old and new pseudonyms, defeating the purpose of the change.

Fig2. An improvement over location privacy using pseudonyms, a trusted proxy and Dummy-Locations In a new approach we can develop a system for location-based service that uses pseudonyms (Fig2). Here too the user accesses the application server through a trusted proxy server. The proxy provides a pseudonym to the user after authenticating it. The user gets the ID from the proxy server administrator on request. User communicates with the application through the proxy server (Fig2). The users change pseudonyms frequently, even while they are being tracked: users adopt a series of new, unused pseudonyms for each application with which they interact [1]. Proxy has to maintain a table for the pseudonym and the corresponding user ID so that it can redirect the response from the application to appropriate user. Here too the application is aware of the location and the request from the user but doesn’t know her identity. Now one more factor is added to make the system more secure. This can be called location anonymity approach (Fig2) [6]. Here the user sends a few dummy locations with its actual location to the application and requests some service. The application server responds with solutions (services) for all locations. The user probability of loss of privacy will decrease chooses the solution for actual location. Hence, Fig6 shows how dummy locations are chosen for multiple locations requests (queries). AL is the actual location. L1 and L2 are chosen to make the application wonder that which one is the correct location of the user. There can be different ways of choosing the dummy locations (DL).

Now we will discuss the proposed method in detail. For illustration we take 5X8 grid area.

Fig3. Path followed by a user and its dummies.

Fig3 shows the user T and its dummies i.e. d1, d2. T starts from (3, 8) goes to (2, 1). Dummy d1 starts at (1, 7) and reaches (5, 6). Dummy d2 starts at (5, 1) and its destination in (1, 3). The user and dummies are guided by application server. When original user moves one block the dummy also moves one block. They continuously send there location to the server through trusted proxy. Users have temporary pseudonyms. They change their pseudonyms according to some algorithm while moving. The application server is in no position to identify them. But it is easy to identify relationship between temp pseudonyms of same user. So complete path followed by a user can be found. Any additional information from some other source can be used to identify the actual user. If user identity is found location privacy is lost. So we have to devise some other method for maintaining location privacy of the user.

Fig4. Path followed by T and its dummies till the change of pseudonym.


7

Now, suppose T, Tx, Ty are three users accessing application server through trusted proxy.

Fig4 shows a user T. The bold line shows the original path followed by the user. Dummies d1 and d2 follow path shown by thin lines. The red squares show the positions when the user changes pseudonym.

Fig5. Path followed by Tx and its dummies till the change of pseudonym.

Fig5 shows a user Tx. The bold line shows the original path followed by the user. Dummies d1 and d2 follow path shown by thin lines. The red squares show the positions when the user changes pseudonym.

Fig6. Path followed by Ty and its dummies till the change of pseudonym.

Fig6 shows a user Ty. The bold line shows the original path followed by the user. Dummies d1 and d2 follow path shown by thin lines. The red squares show the positions when the user changes pseudonym. Important: T, Tx, Ty change their pseudonyms at the same time.

Fig7. Path followed by some user and its dummies after change of pseudonym.

Fig7 shows the path followed by some user (T or Tx or Ty) and its two dummies. It is very difficult to guess who among the three this user is. Why? It will become clear after seeing next fig6.

Fig8. Complete path followed by T and its dummies before and after change of pseudonym.

Fig8 completes fig5. A in fig5 is same user that is T. B and C are dummies. In fig2 we can see that T changes pseudonym when it leaves block (2, 5) and moves to (3, 4). In fig3 we can see that Tx changes pseudonym when it leaves block (1, 3). In fig2 we can see that Ty’s dummy changes pseudonym when it leaves block (4, 2). Now when T appears with new pseudonym the proxy server initializes its dummies i.e. Dx (or B), Dy (or C) in a tricky fashion. Dx is chosen one move away from block (1, 3) where Tx changes pseudonym. Dy is chosen one move away from block (4, 1) where d1 (one of the dummy) of T changes pseudonym. So T appears as A with dummies B, C. What happens next is shown by fig5. So, we can relate it to all three users i.e. T, Tx, Ty. Probability of finding that A is T equals to 1/3.

If there are K dummies and K users change their pseudonym concurrently using this trick to initialize dummies, the probability becomes 1/K. If in time T, users change their pseudonyms N times then there are


8

KN different paths possible for every user in time T. For time 2T, possible paths are K2N. Now even if one has additional information, she is not able to break into the privacy of user. Value of K can be chosen taking in account the present computational speeds of machines (the attacker, the proxy server and the application server) for obvious reasons.

IV. RELATED WORK

The field of anonymous communication originated with Chaums mix networks [2] and the dining cryptographer algorithm [3]. In [2], he proposed an untraceable communication system called the mix that used a mail system, digital signatures. In [3], he also proposed intractability between sender and recipient and the origin of Anonymity Set. A prominent work on location privacy is Mix Zones [1], which is similar to mix networks. In Mix Zones, infrastructure provides an anonymous service using pseudonyms that collects and reorders messages from users within a mix zone to confuse observers. There must be enough users in the mix zone for effective location privacy. Gruteser and Grunwald proposed another mechanism called spatial and temporal cloaking [4] that conceals a user within a group of k people, called k-anonymous, which originated from k-anonymity [5]. To achieve k-anonymous, spatial or temporal accuracy of location information is reduced. But when there are few people in a small area, the accuracy of location information is too low to use for location based services.

V. CONCLUSION

In this paper, we proposed a new technique for location-based services to protect location privacy using dummies. The client creates dummy position data that is sent to the application server with its original position. There is a proxy server in between that anonymizes the user by providing it a pseudonym for communication with the application server. Pseudonyms are changed frequently. If there are K dummies and K users change their pseudonym concurrently and in time T, users change their pseudonyms N times then there are KN different paths possible for every user. This makes this method a good contender for being used in pervasive computing environment. In future we will try to find out its actual effectiveness by simulating it in real-like scenario. Also we need to find out how should the dummies move (what imaginary path should dummies take) so they can’t be distinguished from real users.

VI. REFERENCES

[1] Alastair R. Beresford and Frank Stajano, “Location Privacy in Pervasive Computing”, IEEE Pervasive computing, January–March 2003, pp. 46-55.

[2] D. Chaum. Untraceable electronic mail, return addresses, and digital pseudonyms. Communications ofrhe ACM, 4(2), February 1981.

[3] D. Chaum. The dining cryptographers problem: Uncondi- tional sender and recipient untraceability, Journal of Cryproiagy, 1:65-75, 1988.

[4] M. Gruteser and D. Grunwald. Anonymous usage of location-based services through spatial and temporal cloaking. In Proceedings of the First Inremafional Conference on Mobile Systems, Applications, and Services, pages 3 1-42, 2003.

[5] P. Samarati and L. Sweeney. Protecting privacy when disclosing information: k-anonymity and its enforcement through generalization and suppression. Technical report, 1998.

[6] Tun-Hao You, Wen-Chih Peng, Wang-Chien Lee, “Protecting Moving Trajectories with Dummies”, 2007 www.cs.nctu.edu.tw.


9

Performance of Mortar Incorporating RHA Under Elevated Temperature

Muhammad Harunur Rashid1, Md. Keramat Ali Molla2, Tarif Uddin Ahmed3

1 Muhammad Harunur Rashid, PhD Student, 2 Md. Keramat Ali Molla. Professor, Department of Civil Engineering, Khulna University of Engineering & Technology, Khulna, Bangladesh

[email protected] 3 Tarif Uddin Ahmed. Professor, Department of Civil Engineering, Rajshahi University of Engineering & Technology,

Rajshahi, Bangladesh.

Abstract-- Mortar is one of the important components of concrete. At the time of fire in structure, the mor tar faces the major problem and finally becomes debonded or spalled. At the time of fire different techniques were applied to control it. Applying of water jet is one of the techniques. The performance of mortar subjected to high temperature was examined in this work. Six series o f cubical 5cm × 5cm × 5cm mortar specimens were cast from OPC with partial replacement (10, 15, 20, 25 & 30%) of OPC by Rice Husk Ash (RHA). ASTM graded sand was used as fine aggregate. After 90 days curing in lab oratory condition these specimens were heated in electric f urnace to 200, 300, 400, 500 and 700 oC for 30 minutes. After burning the specimens were removed from the furnace and kept in normal environment and quenched with wa ter for 10 seconds and then kept in ambient temperature to loose the heat. After one day from heating samples were then tested. In this work it was observed that the strength was higher than the controlled sample up to 20% replacement levels under elevated temperature.

Keywords: Rice Husk Ash, Temperature, OPC, quenched.

I. INTRODUCTION

The purpose of this work is to study the performance of mortar incorporation with rice husk ash under the influence of high temperature on microstructure and mechanical properties. Fire is one of the most serious risks and creates serious problem in any structure. Any building may be caught by fire at any time. Most of the structural materials exposed to firing at high temperature are weakened which is not visible. Hence the study on the performance of structural elements under fire has become more and more important.

An increasing number of research works in this field are being done. A very important portion of the fire on cofcrete is degradation of mechanical properties of

concrete. This problem has been studied since 1950’s [1,2,3]. Nijland and Larbi reported that concrete heated by fire might result in a variety of structural failure such as cracking, spalling, debonding of aggregate from mortar, expansion, loss of strength and mineralogical/chemical changes [4].

The researchers stated that when a concrete structure is exposed to fire, differential expansion and contraction of various components and constituents within the concrete take place. St. John et al. stated that with regard to cement paste, evaporation, dissolution, dehydration and dissociation of ettringite, gypsum, calcium hydroxide, calcium carbonate and other phases such as calcium silicate hydrates in the cement paste may be found due to the fire effect [5]. Hansen and Ericsson studied the effects of temperature change between room temperature and 100 oC on the behavior of cement paste, mortar and normal concrete under load [6]. Results of their investigation show that cement paste and mortal beams deflect excessively when heated after application of load. Castillo reported the effect of transient high temperature on the uniaxial compressive strength of high-strength normal concrete. The temperatures studied varied from 100 to 800 oC [7] .

Several studies concerning the fire performance of Light, Normal and high strength concrete were performed, but the mortar with Rice Husk Ash is still not investigated. The references [8-10] have measured and made comments on the residual strength of the concrete under fire, claiming that concrete is one of the most fire resistant material and light weight aggregate concrete has a better fire resisting property than normal density concrete. Sarshar examined the degradation of compressive strength of cement paste specimens had been produced from Ordinary Portland Cement (OPC)


10

[11]. The specimens were cylindrical in shape and height to diameter ratio was 1(one). All the tested specimens’ were heated up to 300 or 520 oC and then cooled in different ways. Sarshar stated that specimens cooled with water shows much greater strength reduction than those cooled slowly.

Cement mortar and coarse aggregate are the main composition in concrete and the mortars are more vulnerable under high temperature. The use of RHA in mortar and concrete as partial replacement of cement has been extensively investigated in recent years. In this research work traditional cement mortar made from OPC and modified cement mortar made from OPC and Rice Husk Ash (RHA) were used.

II. EXPERIMENTAL WORKS

A. Materials

The objective of the experimental program was to investigste the strength properties of cement mortar in presence of RHA as a supplementary material of OPC. The materials used were ordinary Portland cement (OPC) complying with ASTM Type I, ASTM graded sand as fine aggregates and RHA obtained from laboratory combustion process, which is quite similar to traditional uncontrolled combustion process available in rural Bangladesh. The chemical composition of RHA is shown in Table 1.

The mix proportion of cement mortar is 1: 2.75 by weight of material for all samples. Controlled sample is of zero percent of RHA and is designated by A0. Other samples were mixed in different proportions of RHA with OPC cement i.e. OPC cement was partially replaced by RHA. Detail mixtures of mortar with sample ID are shown in Table 2.

TABLE I

Chemical Composition of Rice Husk Ash

Constituents Fe2O3 SiO2 Al2O3 CaO MgO L.O.I.

%Composition 1.28 91.43 0.76 0.91 1.12 3.86

TABLE II

Mixing Proportion for Preparation of Mortar

Mix ID A0 A10 A15 A20 A25 A30

% of OPC 100 90 85 80 75 70

% of RHA 0 10 15 20 25 30

B. Sample Preparation and Testing

After mixing cubical specimen of 5cm × 5cm × 5cm size were prepared for individual % of RHA. For each % of RHA samples were prepared for each range of firing tests. Hence total number of samples including control specimen is 36 for individual mixing ID. Before demoulding all samples were kept in a moist place for 24 hours. Then samples were left for curing in a water tank at 25±2 oC temperature up to the time of test. Before starting the firing tests all samples were preconditioned. It was done by two steps. First keeping those in open air for 12 hours and then heating the sample at 100 oC for 12 hours and after this, samples were taken off from oven and cooled down in open air for 12 hours. Afterwards the specimens were taken for firing tests in an electric muffle furnace. The internal chamber of furnace was 30cm × 30cm × 30cm and the maximum heating capacity was 12000C. Five series of firing tests were done at different heating ranges which are 200, 300, 400, 500 & 700 oC respectively. Firing of samples shown in Fig. 1 in each temperature was maintained for 30 minutes. A pliers made with mild steel was used to collect the heated samples from the furnace and electric switch was off before this. After removing the specimens from the furnace, they were cooled in two different ways. Three of six specimens were cooled down freely in open air at room where the temperature was 29 to 32oC. Remaining three specimens of each type were quenched with water at 30 oC for about 10 seconds. After this they were also cooled down freely in the room temperature. At the time of quenching with water huge amount of vapor was produced which is shown in Fig. 2.


11

Figure 1: Removal of sample from Muffle Furnace

Figure 2: Quenched with water

The heating rate in the furnace shows quite similar to the ASTM E119 up to 700 0C. The maximum test temperature was 700oC and the temperature increasing rate up to this was followed according to ASTM standard. The temperature within the furnace was measured by two different thermocouples at different places and determined the final temperature by mathematical average of these two. The Standard and developed temperature in furnace is shown in Fig. 3.

Fig.3: Time- Temperature Curve (tested and ASTM standard)

The initial temperature of the furnace in this test was maintained at 40oC but in ASTM standard it is 20oC. At 5 and 10 minutes the Standard temperature is 538 oC and 704oC and these temperatures observed in the furnace at 4.96 and 9.45 minutes respectively. So the tested temperature was developed in furnace is very close to the ASTM standard. Samples were kept in the furnace for 30 minutes and in that time the furnace was on and maintained the last temperature. After this the samples were moved from the furnace and followed two different ways (described earlier) of cooling and then tests were performed.

III. RESULTS & DISCUSSIONS

The compressive strength test of mortar sample was carried out next day after heating. All samples were tested in universal testing machine for its strength properties. At 90 day age of the sample, control specimen shows higher strength than 25% and 30 % replacement levels and at 10%, 15% and 20% replacement of OPC by RHA samples shows higher strength than the controlled sample. In this case all the samples were tested at ambient temperature and information are given in Table 3. In case of elevated

0

200

400

600

800

1000

1200

0 50 100 150 200

Time in Min.

Tem

per

ature

in 0 C

Tasted

ASTM


12

temperature samples exhibited different strength when it compared with the ambient strength. At 200 0C the controlled sample showed higher strength than any other samples when it is cooled in open air after heating.

TABLE III

Strength variations due to Elevated Temperature and cooled in open air.

At 300 0C temperature controlled samples show lower values than 200 0C strength. Samples having 10%, 15% and 20% RHA exhibit increasing trend in strength up to 200 0C temperature. Samples are in all temperature exhibit lower strength than the controlled sample when the Ordinary Portland Cement was replaced by 30% RHA. At 700 0C controlled sample exhibits comparatively more strength than all other replacement levels. Controlled sample shows 66.7% and 62.6%, 63.7%, 56.9% strength were observed for 10%, 20% and 30% replacement levels respectively. At higher temperature beyond 500 0C samples with RHA exhibits lower performance than the controlled sample and up to 500 0C samples with RHA exhibit better performance than the controlled one. These results are plotted in Fig. 4.

0

10

20

30

40

50

60

A0 A10 A15 A20 A25 A30Sample ID

Stre

ngth

, N/S

q.m

m

32 C 200 C

300 C 400 C

500 C 700 C

(a): Temperature Verses Strength

0

20

40

60

80

100

120

140

160


Stre

ngth

in %

wrt

con

trol

led

Sam

ple

32 C 200 C 300 C

400 C 500 C 700 C

(b) Percentage of Strength Variation regarding to Control Sample at 32 0C

Sample ID

Temperature

320 C 2000 C 3000 C 4000 C 5000 C 7000 C

Strength MPa

A-0 35.8 44.5 42.0 28.2 31.9 23.8

A-10 36.5 42.3 47.8 40.6 32.3 22.4

A-15 37.7 39.7 48.3 43.5 36.5 21.0

A-20 38.1 40.9 46.4 41.1 38.4 22.8

A-25 35.3 43.6 39.2 34.0 29.8 21.1

A-30 32.5 35.0 31.1 26.6 24.6 20.4


13

Fig. 4: Sample cooled in open air after Elevated Temperature

In case of samples quenched with water for 10 seconds immediately after removing from the furnace, mortar strength was dropped in most of the cases with respect to the samples cooled in open air only and the results are shown in Table 4. At 200 0C temperature, all the samples have lower strength except 10% replacement level followed by the controlled sample. At 300 0C temperature, sample having 10, 15 and 20% replacement levels shows higher strength than the strength at ambient temperature. This observation is similar to the samples which were cooled in open air only after heating. It was observed in Fig. 5 that at 400 0C, the compressive strength of samples with 15 and 20% replacement levels is still in increasing trend whereas the controlled sample lost 16.2% strength.

TABLE IV

Strength variation due to Elevated Temperature and quenched with water.

In Fig. 5 it is observed that at 500 0C temperature, controlled sample lost 30.9% strength and sample with 20% RHA lost only 7.6% compressive strength when the samples were quenched with water. And samples having 10 and 15% RHA also shows higher results than the controlled sample. Mortar samples at 700 0C temperature, with and without inclusion of RHA, exhibits lower results than any other temperature series. At this level controlled sample lost 45.9% strength regarding

the strength without elevated temperature. Sample with 10, 15, 20, 25 and 30% RHA lost 43.6, 48.5, 51.5, 59.8 and 61.2% strength correspondingly.

0

10

20

30

40

50

A0 A10 A15 A20 A25 A30

Sample IDSt

reng

th, N

/sq.

mm

32 C 200 C

300 C 400 C

500 C 700 C

0

20

40

60

80

100

120

140

160


Str

engt

h %

wrt

con

trol

led

Sam

ple

32 C 200 C 300 C400 C 500 C 700 C

Sample ID

Temperature

320 C 2000 C 3000 C 4000 C 5000 C 7000 C

Strength MPa

A-0 35.8 40.9 37.3 30.0 24.7 19.3

A-10 36.5 42.9 41.5 34.3 26.5 20.6

A-15 37.7 38.8 40.2 40.1 31.2 19.4

A-20 38.1 35.4 43.0 41.9 33.1 18.5

A-25 35.3 40.9 34.7 28.5 24.3 14.2

A-30 32.5 31.1 28.9 30.1 20.6 12.6


14

Figure 5: Sample quenched with water after Elevated Temperature and then cooled in open air.

The strength of all samples for 200 and 300 oCC was plotted in the graph shown in Figure 6. It was observed that the strength of controlled specimen at 200 oC temperatures was 124.4% and 114.3% higher at open air and quenched with water cooling condition respectively at ambient temperature. In case of open air cooling the controlled sample exhibits highest strength among all the samples and in case of quenched with water highest value shown in 10% replacement levels and which 120.1% is. At this temperature samples having 10% replacement levels shows higher strength when quenched with water than open air cooling condition. The Compressive strength of mortar increased up to a maximum of 135% for sample of 15 percent replacement levels in case of open air cooling and 120% for sample having 20% RHA in case of quenched with water when it was heated to up to 300 oC for controlled sample. All samples when quenched with water exhibits lower strength than the open air cooling after at 300 oC temperatures for 30 minutes.

0

10

20

30

40

50


Stre

ngth

, N/S

q.m

m

Open Air Cooling

Quenched and then air Cooling

(a) Samples at 200 0C

0

10

20

30

40

50


Stre

ngth

, N/S

q.m

m

Open Air Cooling

Quenched and the air cooling

(b) Samples at 300 0C

Figure 6: variation in strength due to cooling system at Temperature of 200 & 300 0C

In case of open air and quenched with water the controlled sample at 400 0C exhibits lower strength than in case of ambient temperature. At this temperature samples with 10%, 15% and 20% RHA still showed increasing nature in its compressive strength when cooled in open air on the other hand when samples are quenched with water 15% and 20% RHA samples gain highest strength. All other samples at this elevated temperature showed lower strength when compared to the strength of ambient temperature. At 500 0C temperature all samples were dropping their strength in both case of cooling system compared to the ambient result. In the case of quenched with water the compressive strength shows 14% to 20% lower values than the samples directly cooled in open air. Figure 7 shows the detailed information about the samples facing 400oC and 500oC temperatures and cooled in open air and quenched with water.


15

0

10

20

30

40

50


Stre

ngth

, N/S

q.m

m

Open Air Cooling

Quenched and then Air cooling

(a) Samples at 400 0C

0

10

20

30

40

50


Stre

ngth

, N/S

q.m

m

Open Air Cooling

Quenched and then Air Cooling

(b) Samples at 500 0C

Fig.6: variation in strength due to cooling system at Temperature of 400 & 500 0C

Finally the strength was reduced to 67% when firing was performed at 7000C and cooled in open air system for controlled samples. Samples having 10% and 15% RHA exhibits only 4.9% and 4.3% loss of strength when quenched with water in compared to the open air cooling system. This result is shown in Figure 7.

0

10

20

30

40

50


Stre

ngth

, N/S

q.m

m

Open Air Cooling

Quenched and then Air Cooling

(e) Samples at 700 0C

Fig.7: variation in strength due to cooling system at Temperature 700 0C

The degradation of strength at elevated temperature is quite similar to the observation by Sarshar and Nassif et.al. on concrete cylinder [11,12]. Specimens were heated up to 500 oC and cooled by quenching with water by Bazant mentioned that the compressive strength is much lower than specimens not quenched but allow cooling slowly [13].

IV. CONCLUSIONS:

The following conclusions can be drawn from this research work

1. The addition of RHA as cement replacing material is quite satisfactory when the samples are in elevated temperatures.

2. At higher temperature the performance of mortar incorporating Rice Husk Ash shows better behavior on compressive strength than OPC mortar.

3. At 200 oC the controlled sample shows highest strength than any other samples when cooled in open air and 10% replacement levels exhibits highest compressive strength among all the samples when the samples were quenched with water and then cooled in open air

4. Samples having 20% RHA shows better performance up to 500 oC temperature for the controlled as well as other samples having different percentages of RHA.


16

5. The short time quenched with water of the heated specimens has produced significant effect on the compressive strength of mortar.

6. At 20% replacement level of OPC by RHA is accepted under elevated temperature.

V. REFERENCES

[1] Malhotra, H.L. The effect of temperature on the compressive strength of concrete. Magazine of concrete research, Vol 8 No 23. Aug 1956, pp 85-94.

[2] Zoldners, N.G. Effect of high temperature on concretes incorporating different aggregates. American Society of Testing Materials, 60/1960, pp1087-1108.

[3] Brams, J. A., Compressive strength of concrete at temperature to 16000 F; ACI publication SP25 Paper SP25-2; American Concrete Institute, Detroit, 1971.

[4] Nijland T. G. and Larbi, J. A. “Unraveling the temperature distribution in fire-damaged concrete by means of PFM microscopy: Outline of the approach and review of potentially useful reactions”, Heron, 2001, 46, 253-264.

[5] St. John, D. A. Poole A. W. and Sims, I. “Concrete Petrography: A Handbook of Investigative Techniques”, Arnold, London, 1998.

[6] Hansen, T. C. and. Ericsson, L “Temperature change effect on behaviour of cement paste, mortar, and concrete under load”, J. Am. Concrete Insti., 1966, 63, 489-504.

[7] Castillo, C., “Effect of transient high temperature on high-strength concrete”, MS. Thesis, 1987, Rice University, USA

[8] Lie, T.T., Celikkol, B.; 1990:” Method to Calculate the Fire Resistance of Circular Reinforcement Concrete Columns”. ACI Materials Journal, vol. 88, No.2,pp 84-91.

[9] Schneider, U; 1985, “ Behavior of Concrete at High Temperatures”. RILEM committee 44-PHT, Kassel Tyskland 1985.

[10] Opheim, E., 1995 ; “ Residual strength of fire exposed structural elements.” Report 6.4, High Strength Concrete phase3. SINTEF-report nr STF70 A95205, Trondheim. Pp 29.

[11] Sarshar, R. Effect of Elevated Temperatures on the strength of different cement pastes and concrete. PhD thesis, University of London, 1989.

[12] Nassif, A. Y.; Rigden, S.; Burley, E. 1999, “ the effect of rapid cooling by water quenching on the stiffness properties of fire-damaged concrete, Magazine of Concrete Research, No 4, Aug, p. 155-161.

[13] Bazant, Z. P.; Kaplan, M.F.; 1996, “Concrete at High Temperatures, Material Properties and Mathematical Models: Longman Group Limited.


17

Implementation of an Onboard Visual Tracking System with Small Unmanned

Aerial Vehicle (UAV) Ashraf Qadir1, William Semke 2, Jeremiah Neubert3

Mechanical Engineering Department, University of North Dakota Grand Forks, North Dakota-58201, USA

1 [email protected] 2 [email protected]

3 [email protected] Abstract—This paper presents a visual tracking system

that is capable of running real time on-board a sma ll UAV (Unmanned Aerial Vehicle). The tracking system is computationally efficient and invariant to lighting changes and rotation of the object or the camera. Detection and tracking is autonomously carried out on the payload computer and there are two different methods for cr eation of the image patches. The first method starts detec ting and tracking using a stored image patch created pri or to flight with previous flight data. The second method allows the operator on the ground to select the interest o bject for the UAV to track. The tracking system is capable of re-detecting the object of interest in the events of t racking failure. Performance of the tracking system was ver ified both in the lab and during actual flights of the UA V. Results show that the system can run on-board and t rack a diverse set of objects in real time. Keywords: Unmanned Aerial Vehicle (UAV), Visual Tracking, Kalman filter, Zero Mean Normalized Cross Correlation (ZMNCC), Image Warping, Three axes Gimbal.

I. INTRODUCTION Visual tracking has been an active research topic due

to its potential in wide range of applications in robotics and autonomous systems like vision based control [1], surveillance [2], augmented reality [3], visual reconstruction etc. One major research area is visual tracking with small UAV and its applications. Vision based tracking system has applications in vision based navigation [4], sense and avoid system [5], traffic monitoring [6], search and rescue, etc. They are particularly effective in GPS denied environment and tracking non-cooperative targets. For example, vision based tracking system can be used in tracking a car in an urban environment or rescue a person in woods.

Developments in auto pilot technologies and reduced

cost has increased the potential applications of small unmanned aircraft systems. A considerable amount of

work has been done on visual tracking with UAV and its application in vision based navigation, autonomous control and sense and avoid system for UAV [7, 8]. However, the major challenge in visual tracking with small UAVs is the computation cost of the tracking algorithm. Small UAVs have limited payload capacity and majority of the visual tracking systems are computationally expensive [11]. As a result most of tracking systems use powerful ground stations to run the tracking algorithm and the target position information is then sent to the UAV. The tracking system described in [9] uses an on-board camera to capture the video and sends the data to a ground computer that runs off-the-shelf (COTS) image processing software. The target information is extracted and guidance commands are then sent back to the UAV. These methods depend heavily on the communication between the UAV and the ground computer, thus a communication loss between the UAV and ground computer results in tracking failure. A linear parametrically varying (LPV) filter based motion estimation algorithm in their tracking system was proposed in [10] where the target-loss events due to communication interruption have been modeled as brief instabilities. However the algorithm shows a degradation of performance in presence of target loss events.

An on-board visual tracking system eliminates the

dependency on the communication with the ground station and makes the system less prone to failure. However small unmanned aircraft systems have limited payload capacity and power budget. As a result an onboard visual tracking system requires a tracking algorithm which is robust and computationally efficient. An on-board visual tracking system for UAV control was described in [11]. The tracking system used a scale invariant feature transform (SIFT) algorithm for detecting salient points at every processed frame for visual referencing. Test results show satisfactory matching but at a rate not sufficient for real time tracking and they


18

found tracking speed depends heavily on the size of the search window.

This paper presents a real-time visual tracking system developed to run onboard a small UAV. The algorithm is based on a similar system presented in [12]. Testing of the system was limited to simulated flight data. Based on the initial results, the system was changed to reduce the computational expense of the algorithm and improve its robustness. This improved system was evaluated using laboratory experiments and actual flight tests. The results show that the system is robust and capable of real-time operation on a small UAV.

The remainder of the paper is organized as follows:

The basic architecture of the tracking system is described in Section (2). Section (3) describes the hardware and software implementation of the tracking system. Testing and results of the system are presented in Section (4). Future works with concluding remarks are presented in Section (5).

II. METHODOLOGY

This section presents our tracking system, which includes: object detection using template matching, image warping, kalman filtering and camera actuation.

A Zero Mean Normalized Cross Correlation (ZMNCC)

based template matching method was used for object detection as the object appears small from the UAV flying at 600 feet and its scale does not change significantly. Rotation invariance was achieved by creating a set of templates with 10 degrees interval using image warping and comparing the templates with the image frames. A kalman filter [13, 14] was used to make the system computationally efficient by limiting the search region in the image frame.

A. Template Matching With Zero Mean Normalized Cross Correlation

A large number of tracking methods have been proposed for visual tracking. These methods vary in object representation and detection, and choice of a particular method depends on the application. A comprehensive description of different approaches for object representation and detection can be found in [15]. A vehicle on the ground from the UAV flying around 600 feet appears very small with the resolution as small as 20x20 pixels. Therefore it is difficult to extract enough features for feature based object detection. On the other hand, template matching techniques have been proved to be effective for recognition and classifying small objects.

The tracking system uses zero mean normalized

cross correlation method to detect the object of interest

in the video frames captured by the payload on-board the UAV. A small image patch of the interest object is used as the template and the object is detected comparing the template and the image using zero mean normalized cross correlation. The template is slid over the image and correlation coefficient is calculated to detect the position of the template in the image frame. A detailed description of normalized form of cross correlation can be found in [16, 17] where they also proposed fast algorithms to calculate zero mean normalized cross correlation coefficient. The Fast Normalized Cross Correlation Coefficient equation is described as

[ ][ ]

[ ] [ ]{ } 2/122

,

,

),(),(

),(),(

∑

∑

−−−−

−−−−=

tvyuxtfyxf

tvyuxtfyxfC

vu

vu

(1) where C is the Zero Mean Normalized Cross

Correlation coefficient, t is the mean intensity value of

the template and vuf , is the mean intensity value of the

image f(x,y) in the region under the template. Subtracting the mean and normalizing the image and template make the correlation coefficient value ranges from -1 to +1. A best match between the image region and the template results in a coefficient value of +1 and -1 means a complete disagreement between the template and image. Different threshold values for the correlation coefficient (C) have been used in the algorithm during Experimentation with the previous flight videos. Results show that a correlation coefficient value of 0.9 or higher gives a true match between the image and template.

B. Image Warping Zero mean normalized cross correlation (ZMNCC)

makes the tracking system invariant to intensity changes in the image sequences. However, template matching with ZMNCC detects object where there is only translation or small changes of the interest object shape or orientation. An object viewed from the UAV flying 600/700 feet appears small and its shape does not change much. But both the object and/or camera have rotation when tracking with a UAV. Image warping has been used to make the tracking system rotation invariant. A set of 36 templates have been generated from the original image patch with 10 degrees interval using image warping to accommodate full 360 degree rotation of the object. Templates are then compared with the image to detect the object of interest.

Image warping can be defined as mapping a position (x, y) in the source image to the position ( yx ′′, ) in the destination image [18]. If a position in the source 2D image expressed in homogeneous coordinates as


19

Tyxx ]1,,[= , and its corresponding position in the

destination image as Tyxx ]1,,[ ′′=′ in homogeneous coordinates, then the mapping can be described as

Hxx =′ (2) where H denotes the Transformation matrix. For pure rotation the transformation is expressed as

−=

′′

1100

0cossin

0sincos

1

y

x

y

x

αααα

(3) However not all 36 image templates are compared

with the image to find the match every frame. The algorithm starts with comparing the first frame image with the templates. When a match is found, the algorithm leaves the cross correlation part of the program and the template number is recorded. In the next frame, the algorithm starts two patches previous to the patch number that was recorded in the previous iteration. However, the maximum number of templates compared in one frame is 7. If the algorithm does not find a match while comparing all 7 patches, it goes to the next frame. Now the algorithm starts at one patch after the one it started at in the previous frame. For example, once the system has the template, it generates 36 image templates with 10 degrees rotation. The algorithm starts cross correlation with template number 1 in the first frame. Say it finds the match at template number 4. The algorithm leaves the cross correlation and goes to the next frame. In the next frame, the algorithm will start with template number 2 and compare up to frame number 8. If there is no match in this frame, the algorithm goes to the next frame and starts the template matching with patch number 3.

C. Kalman Filtering

An extended Kalman filter has been used to make the tracking system computationally efficient and capable of running real time on-board the UAV. Moving the template over the entire source image and compute the correlation coefficient at every position is computationally expensive and consumes a lot of time. Predicting the position of the interest object in the next frame and searching only a region around the predicted position make the system computationally efficient. The position of the interest object is predicted using the motion model of the Kalman filter. Then a search window is generated around the predicted position using the process covariance matrix of the Kalman filter. The position is then estimated using the measurement from the object detection and the process covariance matrix is updated. In the absence of any detection, the state is

not updated, another prediction is made and the detection process repeated. The covariance matrix gets bigger which in turn makes the search window bigger. Two governing equations for extended Kalman filtering are:

),,( 111 −−−= kkkk wuxfx (4)

),( kkk vxhz = (5)

Equation (4) is the non-linear stochastic difference

equation where kx is the state of the system, k is the

time stamp, the function f describes the process model of the system and w is the process noise. Equation (5)

relates the measurements kz to the state kx of the

system at the time stamp k. An extended Kalman filter is used for the vision

tracking of the complex trajectory (change of acceleration) of the object [19]. Assuming a fixed velocity model, the state of the system is described as

=

k

kk x

xX

&v

r

, (6)

where kX is the system’s state, kxr

is the object’s position

and x&v

is the velocity of the object at time instant k The dynamic system was modeled as:

11 *]exp[ −− ∆= kkk XtxXr

&vv

(7)

The filter is initialized with the following items Process Jacobian: For a 2 degree of freedom system

the process Jacobian

∆∆

=

1000

0100

010

001

t

t

A (8)

Jacobian matrix A is dependent on the time elapsed

between observation k-1 & k and is denoted as t∆ . Process noise covariance matrix is initialized as

∆∆

=

44

33

42

31

00

00

00

00

σσ

tb

tb

ba

ba

Q (9)

where 23

3

1+∆+∆= iii tta σσ , (10)


20

and ii tb σ2

2

1 ∆= (11)

The values of σ are determined using experimentation with the previous flight data. The search windows were generated from the covariance matrices around the predicted position. A value of 0.4 for σ generates a search window big enough to keep the object inside the window.

The Measurement Jacobian is

=

∂∂=

0010

0001

X

hH k v , (12)

Measurement noise kR is projected into state space

using the equation Tkkk VRV . The following measurement

noise matrix here has been used:

=

10

01kR ,

(13) where the elements represents one pixel of uncertainty in object localization in the image frame. The steps involved in using Kalman filtering in our vision tracking system are:

1. Initialization (k=0): In this stage the whole image is searched for the object due we do not know previously the object position. The object is detected in the image

frame and its centre is selected as the initial state 0x̂ at

time 0=k . Process covariance matrix kP̂ is also

initialized.

2. Prediction (k>0): The state of the object −=1ˆkx is

predicted using the motion model of the Kalman filter for the next image frame at time k=1. This position is considered as the center of the search window to find the object.

3. Correction (k>0): In this stage the object is detected within the search window (measurement kz ) and the

state kx̂ and covariance matrix

kP is updated with the

measurement data. Steps 2 and 3 are carried out while the object tracking

runs. The size of the search window is dictated by the noise in the prediction and depends on the process covariance matrix. A small process covariance matrix results in a small search window size and implies that the estimation is trusted more. In the absence of detection, there are no measurements and state is not updated. This results in a larger process error covariance and the search window gets bigger.

D. Actuating the Gimbal

Once detected, the distance between the center of the image and the detected position is computed. The

distance is then converted to motor count and sent to the motion controller for the pan and tilt motion of the gimbal. The gimbal actuates accordingly to keep the detected object at the center of image frame.

III. IMPLEMENTATION

The tracking system was implemented on the SUNDOG (Surveillance by University of North Dakota Observational Gimbal) payload [19, 20] developed by the undergraduate mechanical and electrical students at the University of North Dakota. A customized UAV named “Super Hauler”, owned and operated by the UASE lab, UND houses the SUNDOG payload as well as the Piccolo autopilot [21] with its dedicated control link. The UAV has a wingspan of 144 inches and 120 inches of length. The UAV weighs 48 pounds and has

25 pounds of payload carrying capacity.

Fig. 1. BTE Super Hauler UAV

A. SUNDOG Payload

The payload consists of a PC/104+ form factor based computer- essentially a Linux PC on a single printed circuit board (PCB) with frame grabber, additional octal serial port board and wireless card, a three-axis precision pointing system for an Electro-Optical camera and an Infrared camera. A 2.4 GHz PC/104-plus form factor based wireless card and a RTD PC/104- Plus Dual Channel Frame Grabber is stacked with the computer. A color Sony FCBEX980 camera is mounted on the gimbal. The gimbal has 360 degrees rotation and 30 degrees pan and tilt rotation motion.

Fig. 2. SUNDOG [19, 20] payload. Image shows the three-axis gimbal system with the Electro-Optical (EO) and Infrared (IR) cameras mounted on it.


21

B. Motion Controller

FaulHaber motion controller [22] use pulse-width modulation (PWM) signals to drive the DC servo motors. The drive or amplifier transforms the PWM signal into high amplitude current to turn the motors. They allow for torque control via current regulation. Incremental encoders have been used for position feedback. The controllers are connected via RS232 serial cable to the on-board computer and provide resolution of 100 micro-radians (0.00570).

C. Joystick Control

A joystick control of pointing the gimbal has been implemented for on-line selection of the interest object from the ground. Joystick control allows the ground operator to manually point the gimbal at the target for selection.

IV. EXPERIMENTATION AND RESULTS

A. Testing with Previous Flight Video

The algorithm was tested with video captured in

previous test flights with two objectives in mind. One is to check that our detection algorithm is robust enough to locate objects on the ground and the second one is to tune the algorithm so that it is computationally efficient and capable of real time tracking. The algorithm was tested with different threshold values for the normalized cross correlation coefficient and search window size. The search window size depends on the covariance matrix of the Kalman filter. Process covariance matrices were generated by selecting different values of covariance matrix elements (σ ). Then the system was optimized by generating the search window which is big enough to make sure that the object stays inside the search window in successive frames. An object is selected in one image frame and then tracked in subsequent frames, as shown in the Figure 3. The object is selected in the left image and the template is generated. The red rectangle represents the selection area in the source image frame. The right image shows the object tracking. The red dot in the image represents the detected object which is the centroid of all matching points between the template and the source image. The blue rectangle around the object is the search window generated by the covariance matrix of the Kalman filter while the red rectangle represents the initial position of the object.

Fig. 3. Tracking on the previously captured image frames. The object was selected on the left frame and the right frame shows the object detection.

The recorded data was used to determine the optional

parameter settings for in-flight operation. The initial testing was used to select the threshold for ZMNCC. This was accomplished by tracking several objects in the video sequence with differing threshold values. The selected value, 0.9, was able to successfully match the desired object more than 95% of the time with no false positives.

After the threshold was selected, the video was used

to determine the noise in our process model. This was accomplished by assuming a small amount of uncertainty and attempting tracking. The tracking algorithm was limited to a search region based on predicted uncertainty assuming that the object remains within three standard deviations of its predicted region. If an interest object fell outside the search region, the estimated model uncertainty was increased. The process continued until the uncertainty allowed all the interest objects in the video sequence to be maintained in the predicted search region. Using a model uncertainty matrix with 0.4 on the diagonal produced the desired results.

B. Tracking with Model Cars

A laboratory experiment has been carried out by simulating the flight environment. The objective of the experiment was to verify that the tracking algorithm was robust and computationally efficient. The algorithm was also tuned with the experimental data. A small moving car was used as the target object. The payload was mounted on a stand 3 feet (0.9144 meters) above the ground. The goal of the payload was to track the car on the ground and move its gimbal to keep the car at the center of the image frame. The middle section of the top of the car was selected as the template. The tracking algorithm detected the car and started tracking. To verify the robustness and efficiency of the tracking system, the car was translated, rotated and moved around other objects. Tracking was displayed on the screen in real time as the system was running and intermittent image frames were saved for later analysis purpose.


22

Online selection of an object of interest and sending

this data to the payload was also tested in the lab using a direct serial communication between ground computer and the payload. Video frames were sent over the Ethernet from the payload to the computer used as a ground station. The car was selected on the ground computer screen and the image patch was sent to the payload using the RS232 communication. Upon receiving the image patch, the payload started tracking.

Results of the experiment are shown in Figure 4. The

red dot shows the centroid of the detected positions and the green rectangle around it is the search window generated by the covariance matrix of the Kalman filter. The object appeared with different position and orientation as it was moved and the tracking system was able to detect the car.

Fig. 4. Tracking model car in the lab. The car was translated and rotated to test the performance of the tracking algorithm

The gimbal was actuated to keep the car at the

center of the image frame. The tracking system exhibited its ability to real time track and re-detect objects in the event of tracking failure in one frame. The system was able to predict the state of the tracked object and the search region was able to keep the object inside.

Computation cost of the tracking system was also

verified by computing the frame rate the tracking system was capable of processing. Results show that the system was able to track real time with more than 25 frames per second. Results also show that the tracking speed varies with the patch sizes. Different size patches were selected during the tests and being tracked by the algorithm. Smaller patches resulted in higher tracking speed. Table 1 shows the frame rates for different size patches.

TABLE 1

TRACKING SPEED WITH DIFFERENT PATCH SIZES

Patch Size

(Pixels)

Number of Frames Tracking Speed

(Frames/Sec)

27x28(756) 813 27.83 20x22(440) 246 28.21 38x30(1140) 558 26.99 30x33(990) 533 27.57

C. Actual Flight Tests

Finally multiple flight tests of the complete tracking system were conducted in May-August of 2011. The flight tests validate the tracking algorithm and shows that the tracking system is capable of re-detecting the interest object in the event of tracking failure. The test bed for the tracking system is explained in Fig. 5.

Fig. 5. Flight test bed


23

Fig. 6. Tracking results from actual flight test. The patch is shown at the bottom left corner of the first image frame. The object was detected in one video frame. The camera was then actuated to keep the object at the centre of the image frame. The red dot shows the detection of the object and the blue rectangle is the search region around the object. In the first frame there is no blue rectangle drawn because it is the first frame where the object was detected and the search region was the whole image. Once the object was detected, the covariance matrix of the Kalman filter was updated with the measured data and the search region gets smaller.


24

The UAV was flying at 600 feet altitude and a

repeated flight path over the ground control tent was chosen for the UAV. The video captured by the payload camera was broadcast to the ground computer using 2.4 GHz communication link. On the ground the interest object was selected from the video and sent to the UAV during one pass of the UAV over the object. Once the image patch is received by the payload, the algorithm generates 36 image patches of the interest object with 10 degrees rotation to cover the whole 360 degrees rotation of the object. The object was lost when it went out of the camera view. Once the UAV came over the object again in the next pass, it detected the object and started tracking. Then the gimbal with the camera was moved to keep the object at the center of the image frame. The tracking system successfully tracked four different objects during the flight tests. The results of the flight test have been shown in Fig. 6.

V. CONCLUSION

An on-board visual tracking with small UAV that is capable of tracking real time has been developed. The tracking system locates the object of interest using zero mean normalized cross correlation between object template and source image. The tracking system is invariant to changes in illumination or rotation of the object. The tracking system uses a Kalman filter to estimate the object position and create a search window around the estimated position. Image warping has been used to make the tracking system rotation invariant. The system was implemented on the SUNDOG payload and was tested using several experimentations including a full hardware in the loop test in the lab and actual flight tests. The tracking algorithm has the ability to re-detect and track in the event of loss of tracking. If the object goes out side the search window the tracking fails. The system quickly recovered from this failure by expanding the search area. The system assumes that large scale changes of the tracked object do not occur. This assumption appears valid during actual flight tests. A simple linear projective transformation can be used in the tracking system with the aircraft altitude information to accommodate scaling. Better control of the camera parameters such as gain, and exposure will reduce the blurriness in the images caused by vibration or very fast movement of both the camera and the object.

ACKNOWLEDGMENT

This research was supported in part by Department of Defense contract number FA4861-06-C-C006, “Unmanned Aerial System Remote Sense and Avoid System and Advanced Payload Analysis and Investigation,” and North Dakota Department of Commerce grant entitled ”UND Center of Excellence for

UAV and Simulation Applications.” The authors appreciate the contributions of the Unmanned Aircraft Systems Engineering (UASE) Laboratory team at the University of North Dakota.

REFERENCES [1] E. Malis,. ”Survey of Vision-based robot control,” Technical

Report, INRIA, Sophia Antipolis, France, 2002. [2] B. Coifman, D. Beymer, P. McLauchlan and J. Malik., “A real-

time computer vision system for vehicle tracking and traffic surveillance,” Transportation Research Part C 6, pp. 271–288, 1998.

[3] D. Koller, G. Klinker, E. Rose, D. Breen, R. Whitaker, and M. Tuceryan., “Real-time vision-based camera tracking for augmented reality applications,” ACM Symposium on Virtual Reality, Software and Technology (VRST-97), 1997.

[4] B. Ludington, E. Johnson, and G. Vachtsevanos., “Augmenting UAV Autonomy: Vision-Based Navigation and Target Tracking for Unmanned Aerial Vehicles,” IEEE Robotics and Automation Magazine, vol. 13, issue 3, pp. 63-71, 2006.

[5] J. B. Saunders., “Obstacle Avoidance, Visual Automatic Target Tracking, and Task Allocation for Small Unmanned Air Vehicles,” Department of Electrical and Computer Engineering, Brigham Young University, 2009.

[6] F. Heintz, P. Rudol, and P. Doherty., “From images to traffic Behavior—a UAV tracking and monitoring application,” In Proceedings of the 10th international conference on information fusion, Quebec: ISIF, IEEE, AES, pp. 1–8, 2007.

[7] Z. He, R. V. Iyer, and P. R. Chandler., “Vision-based UAV Flight Control and Obstacle Avoidance,” In Proceedings of IEEE American Control Conference, Minneapolis, MN, pp. 2166 –2170, June 2006.

[8] C´. Teuli`ere, L. Eck, E. Marchand, and N. Gu´enard., “3D model-based tracking for UAV position control,” IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010.

[9] L. Ma, V. Stepanyan, C. Cao, I. Faruque, C. Woolsey, and N. Hovakimyan., “Flight Test Bed for Visual Tracking of small UAVs,” In AIAA Guidance, Navigation, and Control Conf. and Exhibit, AIAA 2006-6609, 2006.

[10] V. N. Dobrokhodov, I. I. Kaminer, K. D. Jones, and R. Ghabcheloo., “Vision-Based Tracking and Motion Estimation for Moving targets using Small UAVs,” In Proceedings of American Control Conference, 2006

[11] I. F. Mandragon., “Visual Model Feature Tracking for UAV Control,” In IEEE International Symposium on Intelligent Signal Processing, WISP 2007, 2007.

[12] A. Qadir, W. Semke, and J. Neubert., “On-board Visual Tracking with Small Unmanned Aircraft Systems,” In AIAA Infotech@Aerospace conference, 2011.

[13] R. E. Kalman., “A new approach to linear filtering and Prediction Problems,” Transaction of ASME, Journal of Basic Engineering (series D) volume 82, pp. 34-45, 1960.

[14] G. Welch, and G. Bishop., “ An Introduction to the Kalman Filter,” Department of Computer Science, University of North Carolina, Chapel Hill, TR95-041.

[15] A. Yilmaz, O. Javed, and M. Shah., “Object tracking: A survey,” ACM Journal of Computing Surveys, vol. 38, no. 4, 2006.

[16] J. P. Lewis., “Fast Normalized Cross-Correlation,” Vision Interface, pp. 120-123 1995.

[17] K. Briechle, and U. D. Haneback., “Template Matching Using Fast Normalized Cross Correlation,” In Proceedings of SPIE AeroSense Symposium volume 4387, 2001.


25

[18] C. A. Glasbey, and K. V. Mardia., “A review of image-warping methods,” Journal of Applied Statistics, volume 25, no. 2, pp.155–172, 1998.

[19] E. Cuevas, D. Zaldivar, and R. Rojas., “Kalman Filter for Vision Tracking,” Freie Universit¨at Berlin, Tech. Rep., B 05-12, 2005.

[20] J. Ranganathan, and W. Semke., “Three-Axis Gimbal

Surveillance Algorithms for Use in Small UAS,” Proceedings of the ASME International Mechanical Engineering Conference and Exposition, IMECE2008-67667, 2008

[21] W. Semke, J. Ranganathan, and M. Buisker., “Active Gimbal Control for Surveillance using Small Unmanned Aircraft Systems,” Proceedings of the International Model analysis Conference (IMAC) XXVI: A Conference and Exposition on Structural Dynamics, 2008

[22] CloudCap Technology, Piccolo-Autopilot. URL: http://cloudcaptech.com

[23] Motion Control System Instruction Manual, Series MCDC 3003/06 S, Faulhaber Inc, URL: www.faulhaber.com


26

IJITCE Oct 2011

Documents

Transcript of IJITCE Oct 2011