Theses & Dissertations

Permanent URI for this collection

https://hdl.handle.net/20.500.14170/209

Browse

Recent Submissions

Now showing 1 - 5 of 361
  • Publication
    DSP implementation of embedded coding for wavelet based image compression on the TMS320C6713
    (Universiti Malaysia Perlis (UniMAP), 2015)
    Iskandar Zulkarnain Shamsudin
    As the new standard compression method for the JPEG 2000, the discrete wavelet transform (DWT) is quickly finding more and more uses in the image compression department. With the numerous and repetitive calculations needed to fmd the coefficients of a large image, speed is critical when implementing a wavelet filter bank to represent the DWT in hardware. The objective o f this thesis is to implement a Embedded Coding for Wavelet Based Image Compression on the TMS320C6713. The reason of this because of the time execution for the embedded coding on computer is take more time to compress one image. TMS320C6713 is a embedded board that provide more fast time execution so that the embedded code can process more fast than the computer.
  • Publication
    Simulation of graphene band structure and fabrication of graphene field effect transistor
    (Universiti Malaysia Perlis (UniMAP), 2018)
    Siti Fazlina Fauzi
    This thesis presents an approach in exploring graphene for digital electronics, as well as analogues. For digital electronic application, this work focused on opening and tuning the monolayer graphene band gap based on the inclusion of dopant atoms, by using AtomixKit Simulator from Quantum Wise. Two sets of simulation have been carried out, that are, to study the graphene band structures based on the effect of asymmetrical bilayer graphene layers and doping the monolayer graphene with boron and Oxygen atoms. Results have shown that monolayer and symmetrical bilayer graphene exhibit zero band gaps, while asymmetrical bilayer graphene provides a band gap of 0.35 eV. The opening of energy band gap in asymmetrical hi- and multi-layer graphene is limited and reduces electron mobility. So another approach to open the band gap without disrupting the high electron mobility in monolayer graphene, is by doping the graphene. This work has shown that the inclusion of dopant atoms in monolayer graphene results in an increase in the band gap, with increased number of dopant atoms. The position of the dopant atoms in monolayer graphene also significantly affects the band gap opening. It is obtained that Oxygen atoms in graphene give rise to higher band gap openings compared to boron atoms. This ability to modify the monolayer graphene band gap can be applied for many digital electronic applications. For analogue application, this research focused on investigating the effect of channel length and gate oxide thickness of graphene field effect transistor (GFET). This GFET fabrication required a good coverage of monolayer graphene. This work has performed an optimization in the graphene transfer process, in which a stamping method using a wooden block is suggested. A proper cleaning process in acetone is also crucial in order to obtain a good layer of graphene before GFET can be fabricated. In this work, the GFET performance is investigated based on the charge carrier mobility taken from both transportation of electrons and holes. The GFETs have been fabricated with two different gate oxides with thicknesses of 63 nm and 83 nm with channel lengths varied from 250 𝛍m to 650 𝛍m. Results have shown that the electron mobility in GFET increases as the channel lengths decreases. The same pattern is observed for hole carrier mobility. However, it is obtained that the electrons and holes mobility decreases as the gate oxide thickness increases. It can be concluded that the GFET channel length and gate oxide thickness play an important role in determining the GFET performance, as the speed of the device increases with shorter channel and thicker gate oxide.
  • Publication
    Synthesis and electrical characterization of Nd-Doped BaTiO₃ for anti-resonance piezoelectric device
    (Universiti Malaysia Perlis (UniMAP), 2017)
    Tuan Amirah Tuan Sulong
    Severe implications by lead usage in microelectronic industry to the environment forced the findings of lead-free ceramics alternatives in recent years. Barium Titanate (BaTi0₃) based ceramics are particularly interesting among the various lead-free ceramics due to its high permittivity, low dielectric loss and exhibit piezoelectric effect. In this study, Neodymium (Nd)-doped BaTiOJ with composition ofBa1-xNdxTi03 at range (0 ~ x ~ 0.13) were prepared via conventional solid state method at high temperature. X-ray diffraction analysis shows that for Nd-doped BaTiOJ, the phase purity was obtained at final heating temperature of 1400 oc for 12 hours. Single phase ofNd-doped Ba TiDJ was observed from 0 ~ x ~ 0.1 0, however, the tetragonal distortion in undoped BaTi0₃ decreased with increasing x. The samples remain tetragonal until x ~ 0.015 and were cubic for 0 . 03 ~ x ~ 0.1 and appear to coalesce at x = 0.13. Maximum permittivity shifted to lower Curie temperature (Tc) from 120 octo 90 oc by doping at tetragonal region and remain linear without maximum at cubic region. For electrical properties, the sample ofx=0.005 exhibit a very high permittivity at room temperature which is 2586, lowest dielectric loss of 0.03, high piezoelectric charge constant about 55 pC.N-1 and high piezoelectric voltage constant about 0. 7 V m.N-1. The novelties in this study was the piezoelectric study was determined at doping concentration of 0 ~ x ~0 . 015 , whereby the piezoelectric properties increased with increasing poling voltage.
      8  2
  • Publication
    Facial recognition vehicle theft prevention system using PCA algorithm
    (Universiti Malaysia Perlis (UniMAP), 2018)
    Mohamed Embaiah Abdulsalam Gerza
    The security system of a vehicle starts with the key system. This is because there is no guarantee that a car invasion will occur. Due to this shortage, the intruders took advantage of the vehicle's theft. Therefore, facial recognition vehicle theft prevention system is developed using the Principle Component Analysis (PCA) algorithm. PCA algorithm consists of four functions which are converts the set of training image to image vectors and normalizes the image vectors. Thirdly, covariance matrix and the Eigenvectors are calculated before the face is recognized. Viola Jones algorithm is also applied in this project to detect the driver's face. This project is implemented using MA TLAB for image processing and Arduino Mega board for tracking system to detect the location of vehicle. The project begins when a person enters the vehicle and sits in the driver's seat, the switch mounted on the driver's seat will activate a hidden camera that focuses on the driver's position. Once the image is taken from an activated camera, the person's face is detected using the Viola Jones algorithm. Executed faces will be identified using the PCA algorithm by comparing with the faces stored in the database in an image format such as the face of the vehicle owner. If comparison of face images is not the same, then warning messages will be sent via the Short Messaging Service (SMS) to the vehicle owners on a regular basis. The SMS contains GPS latitude and longitude information tracked by GPS. This system will light up the green LED when the vehicle intruder is identified and the red LED is lit if the intruder cannot be traced. The design of this system is started by installing the circuit and software development to be the prototype of the project. Finally, vehicle safety systems can be implemented using PCA with the help ofGPS and GSM technology.
  • Publication
    An effective parameters setting for genetic algorithm with local search in job shop scheduling problem
    (Universiti Malaysia Perlis (UniMAP), 2016)
    Inaam Jabbar Kadhim Al-Shammari
    It is well known that GAs are not well suited for fine-tuning structures that are very close to optimal solutions and that it is essential to incorporate local search methods, such as neighborhood search, into GAs. In addition, in solving combinatorial problems such as job-shop-scheduling problem (JSSP) by GAs, efficient parameter settings that provide optimal solutions are usually difficult to construct. This thesis presents an improved approach to Genetic Algorithm (GA) with Local search techniques (LS) in the hybridization Genetic Local Search (GLS) algorithm. GLS attempts to solve the JSSP with the objective of minimizing the makespan value, this can be achieved using effective parameter setting and ideal neighbourhood structure, in GLS algorithm framework. The proposed algorithm consists of three main parts. In the first one, an indirect representation incorporating a schedule builder that performs simple LS is proposed to decode the chromosome into a legal schedule called active schedule. The chromosomes are then decoded into active schedules to significantly increase the probability of obtaining near or optimal solution. In the second part, Multi-Step Crossover Fusion (MSXF) with the intuition of generating a child from the path re-linking technique is used. MSXF, an extended version of LS, utilizes a neighborhood structure and a distance measure in its procedure. In this study, Nowicki and Smutnicki (NS) neighborhood search is used to increase the effectiveness of the algorithm in producing better solutions. By using this type of crossover, a solution or child is generated between both parents using through a search path joining parent solutions. The third part is accomplished by using multi-step mutation fusion as the supplementary search to improve GLS performance. Computational experiments are carried out on the proposed algorithm, which yields better results for most benchmark problems compared with JSSP obtained from the literature. Experiment results indicate that a set of effective GLS parameter settings (population size = 1 0; roulette wheel selection; MSXF crossover; and probability of crossover = 0.5 with mutation operator) exists. The combination of these settings is synergistic, i.e., they collectively enable the procedure to be fast and robust. Meanwhile, NS neighborhood search significantly reduces the total length of the makespan and decrease the computational time. Providing an effective parameter setting facilitates the use of GLS, rendering it suitable for wide-ranging real-time, practical applications in computer science, industries, and all other fields requiring the minimization of costs, delays, etc.
      10  1