M.S. Theses

Permanent URI for this collection

https://digitalarchive.library.bogazici.edu.tr/handle/123456789/11669

Browse

Now showing 1 - 20 of 74

Application of a prototype parallel processing computer for a recurrent neural network model
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 1992., 1992.) Paksoy, Bekir Alper.; Cılız, Kemal.
In this study, Hopfield's binary neural network model is simulated using a software package simulating a bit serial single instruction multiple data (SIMD) mesh array processor, called the Blitzen massively parallel processor (BMPP). First, the parallel algorithms required for the simulation are designed. Then, the parallel implementation of the Hopfield network model has been formulated using these algorithms. The parallel implementation has been analyzed for the speed and the processor utilization. To do this, time complexity of the parallel implementation is derived and compared with the time complexity of a sequential algorithm. Finally, simulation results are analyzed and compared with the analytical derivations. The parallel algorithm described in this study for the simulation of the Hogfield network model on the BMPP architecture achieves a maximum speedup in the order of the square root of the number of processors employed. It is also shown that, for the same algorithm, the maximum possible speedup can be achieved only at a finite number of processors, and the processor utilization decreases as the number of processors is increased.
Identification and control of nonlinear dynamical systems using neural networks
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 1996., 1996.) Efe, Mehmet Önder.; Kaynak, Okyay,
Artificial neural networks opened a new horizon in many research areas. This understanding, also, brought a new way of thinking into the concept of control. The ever-increasing technological demands steered the control engineers to design more sophisticated controllers. In this respect, artificial neural networks were proposed as a new approach because of their massively parallel data processing properties, adaptiveness and powerful mapping capabilities. Especially the learning property of these networks made them extremely attractive. There are various methods that are used for the training of artificial neural networks. Two of them are included in this study. These are, namely, the backpropagation method and the Levenberg-Marquardt optimization technique. The learning time for the former is excessively long especially around the minima since it uses the first order derivatives of the performance function, while in the latter the learning time is very short because of the extra information coming form the second derivatives of the performance function. The computational complexity and the hardware requirements are large for the latter. The identification of nonlinear dynamical systems is a substantial part of the controller training therefore it is included in this work and discussed in the simulation results. The main idea that lies under the procedure is obtaining a regular and mathematically tractable model of the system which is of interest.Based on these two learning methods and system identification, controller design methodology for error backpropagation, inverse control, self-tuning control, model reference adaptive control, self-learning control and dynamical neural unit based control methodologies are explained and excessive simulation results are discussed in this thesis.
Fusion of soft computing techniques in control applications
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Sciences and Engineering, 1997., 1997.) Koku, A. Buğra.; Kaynak, Okyay,; Özkan, Mehmed.
The main objective of this thesis work is the formation of an experimental garden, suitable for the demonstration of classical and advanced control concepts, by , an integration of the computational ca.pabilities and the electromechanical equipment that exist in the Mechatronics Laboratory of Boğaziçi University. An equally important and more scholarly objective has been to construct an experiınental platform on which the use of soft computing methodologies in the corıtrol of nonlinear systems can be studied and experimented with and to actually carry out some experimental investigations in this respect.The thesis has both theoretical and experimental constitueııts, on the practical side it details the electromechanical construction of the X-Garden (stands for eXperimental Garden), the development of the software and the experimental designs, whereas the control algorithms designed and implemented using Soft Computing methodologies make up the theoretical side. The software support developed for the X-Garden as a part of this thesis work can be grouped into two. The first group constitutes the C codes written and the second one is on the experimental designs. Both adaptive and non-adaptive fuzzy controllers are studied both by simulations and by real-time implementations.The thesis carries a didactic objective too. It is intended to provide a first reading on fuzzy and fuzzy-neuro control. The fundamental concepts of fuzzy logic are and fuzzy neuro control are therefore explained in detail, in a clear and tractable manner.The thesis, in some parts, intentionally carries the style of a user's manual, this is to ensure that further users of the X-Garden will have no difficulty in experimenting on the platform constnıcted, being able to find clear explanations and instnıctions in this thesis and its appendices on the use of both the mechanical and the software environments.
Development of a vacuum cleaner mobile robot
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 1998., 1998.) Sardağ, Alp.; Akın, H. Levent.
This study presents the results of the development of an autonomous mobile robot which is a domestic autonomous vacuum cleaner. The robot is designed according to some new concepts established in this field during the last decade. These principles are artificial life, subsumption architecture and other bottom-up methodologies. These ideas have been applied to the complete robot design, spanning from the shape of the robot to the sensors, from the electronics to the software control structure. We have built the robot and programmed it to perform various reactive behaviours. The autonomous vacuum cleaner performs area coverage in a room by combining distinct behaviors. The robot has 11 sensors which are used in the algorithms examined. The robot behaviors allow it to avoid obstacles, follow walls, break out of confined areas, detect floor and find a docking area. The performance tests are done through a dedicated simulator and by observation. The experiments performed here were both in the simulator and real-time it appears that the best area coverage algorithm among those tried consisted of a random walk with a probability of 0.10 for following a wall after each encounter with an obstacle.
Modeling of cardiac images by implicit polynomials for diagnostics
(Thesis (M.S.) - Bogazici University. Institute for Graduate Studies in Science and Engineering, 2000., 2000.) Eskil, Mustafa Taner.; Erçil, Aytül.; Akarun, Lale.
This study aims modeling of left ventricles of human hearts with 2D and 3D implicit polynomials for inspection of heart deformations. Extracting 3D information of left ventricle is made possible by utilization of magnetic resonance (MR) and computerized tomography (CT) technologies. Our research has shown that 4th degree impicit polynomials are adequate both in 2D and 3D for representing the shape of left ventricule and the asymmetric conic shape of the heart for proper functioning. This study consists of a throughout elaboration starting from modeling of human hearts with implicit polynomials, visualization of these 2D and 3D implicit polynomials, calculation of euclidean, affine, and projective invariants and realization of an inspection based on these invariants. Also, a study on any resemblance between the invariants extracted from human and dog hearts is conducted.
Application of variable structure systems theory for training of intelligent systems
(Thesis (M.S.) - Bogazici University. Institue for Graduate Studies in Science and Engineering, 2001., 2001.) Yıldıran, Uğur.; Kaynak, Okyay,
Soft computing architectures with their extensive flexibility and strong mapping capabilities have been widely used for control of nonlinear systems. In this regard, error backpropogation and its derivatives have been the most popular and frequently employed schemes for parameter adjustment of these architectures. However, these schemes bring some serious problems together, like instability of closed loop system and sensitivity to uncertainties, which must be carefully addressed by a system designer. In order to alleviate these problems, recently, Efe has proposed a control strategy in which parameters of intelligent controllers are updated by a continuous-time robust parameters adjustment mechanism in order to robustify and stabilize the closed loop system dynamics. The results obtained for a two link SCARA robot in this study show that the proposed method is successful in achieving the control objectives. In this thesis, the methodology proposed by Efe is investigated for first order nonlinear systems. Based on the results, it has been observed that the time evolution of input-output curves of different structures show similar characteristics. Moreover, a modification is proposed for update mechanism of all architectures in order to prevent unbounded parameter evolution problem which occurs in the original algorithm. Lastly, based on the results for different systems, it has been concluded that the Adaptive Linear Element is the most suitable architecture for the control systems investigated because of its simplicity.
Targeting in chaos using analytically described clusters
(Thesis (M.S.) - Bogazici University. Institue for Graduate Studies in Science and Engineering, 2002., 2002.) Sütçü, Yağız.; Denizhan, Yağmur.
The OGY method provides a simple but powerfbl approach of controlling chaotic dynamics. This method can stabilise inherently unstable equilibrium modes of dissipative chaotic systems under the lack of knowledge about the system equations. However, it has the typical drawback of a long waiting time until the system starting from random initial conditions enters the close neighbourhood of the equilibrium mode to be stabilised, where the controller can be activated. The reduction of this drawback is known under the name of targeting. The Extended Control Regions method is a targeting approach, which can operate under the Iakof knowledge about the system equations by employing local models of the system dynamics extracted from empirical data. The method is based on the idea of identifiing and modelling those regions of the phase space, starting fi-om which the system can be steered to a close neighbourhood of the target within a few steps applying sinall perturbatiotns in the control parameters. So far, the modelling of the system dynamics within these phase space regions have been realised using artificial neural networks. In this study, two different strategies are developed in order to realise the clustered version of the Extended Control Regions method on basis of simple analytical models rather than neural networks. Each cluster obtained the gathered data is analytically described as a hyper-ellipsoid. Subsequently, the analytical models of the clusters are used for targeting purposes by applying small discrete variations in the control parameter. Simulation results on several chaotic systems with single control parameter show that the proposed method can achieve targeting using less memory and computation time than the Clustered Extended Control Regions method on cost of a slower targeting performance.
Improved handling of SMS messages with statistical natural language processing techniques
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2005., 2005.) Yıldırım, Ömer.; Say, Ahmet Celal Cem.; Arslan, Levent M.
The Short Messaging Service (SMS) is built on the ability of mobile telephones tosend and receive text messages. SMS based applications are increasing dramatically day byday in the telecommunications industry. The most common use of SMS is for notifying mobile phone users that they have new voice or fax mail messages waiting. Whenever anew message is dispatched into the mailbox, an alert by SMS informs the user of this fact.The Short Message Service can also be used to deliver a wide range of information tomobile phone users from share prices, match scores, weather, flight information, news headlines, lottery results, jokes. In general, user interaction based SMS services requestsome predefined keywords from the users and respond to them after processing theirmessages.However, most users think that they are communicating not with a machine but withhumans, so they compose misspelled and/or machine specific messages containing more than just the needed keywords. As a result, they receive error messages from the server andgenerally do not continue to use the software after trying two or three times by makingsame mistakes. In this thesis, I introduce a new Short Message Service (SMS) parsing model usingStatistical NLP Techniques, whose aim is to solve the existing SMS user subscriptionproblem of a real software company. To do this, the N-Gram statistical approach will beused.
Two minded prisoners: a skeptic approach to round robin tournaments
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2005., 2005.) Alpar, Orcan.; Egeli, Birgül.
Control of inverted pendulum:| comparison of various strategies for swing up and balancing
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2005., 2005.) Er, Fatih.; Istefanopulos, Yorgo.
In this thesis, two different control schemes are designed for swinging up and three different schemes for balancing of an inverted pendulum, a non-linear electro-mechanical system with open loop unstablity. The controllers which are implemented for experimental purpose are a heuristic control technique using fuzzy logic and an energy control technique for swinging up. In order to balance the pendulum in the upright equilibrium, state feedback control, LQR control which is an optimal control technique and well known traditional control technique PID control are implemented. The control objective is swinging up the pendulum from its initial downward equilibrium to the upward and then balance at this state while it is tracking of desired trajectory. Three different combinations of proposed control schemes are chosen and their performances are compared with each other. Having compared all combinations, it is observed that every controller has advantages and merits for different cases.
Software selection for a liner shipping company using fuzzy logic decision making
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2005., 2005.) Akpınar, Eşref.; Egeli, Birgül.
This thesis mainly focuses on selection of software for managing operations ofEMES Shipping and Transportation Company by the evaluations of scripted scenarios,requests for information (RFI), and demonstrations from software companies. EMES Shipping and Transportation Company, which is a member of Arkas HoldingGroup experiences difficulties and delays in reviewing, approving, controlling andreporting operations between Head Quarter and Agencies due to processes that wereoverburdened with manual tasks, e-mail based workflow, and lack of process information. It is a necessity to select software from the market by evaluating them in the terms ofcompany requirements. This process is a critical issue for future growth andcompetitiveness of EMES.Key operations and activities of the company are obtained from department managers and staff. Gathered information is converted to a generalized request forinformation for the evaluation of the companies and their standard software packages. Fivemajor scripted scenarios are prepared to understand how the software packages handle theoperations of software packages. A demo evaluation document prepared for scripted scenarios to evaluate demonstrations.A fuzzy approach is designed to evaluate companies with RFI scores, developmenttime, and purchase cost variables. Finally an evaluation table and fuzzy based comparisonresults are prepared to show most fitting product to company operations and requirements.
Adaptive fuzzy sliding mode control of a washing machine oscillating system
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2005., 2005.) Tamyaman, Can.; Kaynak, Okyay,
To prevent any mechanical damage, washing machines distribute the load (the laundry)before performing any spins. After some distributive motor movements, there may still besome load that does not lay homogenously on the surface of the tumble. This load may cause undesired oscillations of the system. In cases when the load is not distributedhomogenously, the washing machine senses this unbalanced distribution if it is over athreshold value and reduces the spin speed or does not perform the final spin in thewashing profile. Because the machine is allowed to perform the final spin if the unbalanced distribution measurements are under a certain limit; some undesired oscillations andovershoots in the displacement of the oscillating system of the washing machine is likely totake place. This shortens the mechanical life of the machine. Furthermore, any undesiredmovements of the machine may pose a danger in a household. For these reasons the risk of machine displacement should be calculated and tested on every washing machine at thedesign phase. The machines are either produced with a calculated risk of displacement, orthe spin profile is modified to reduce the risk according to the results of the tests at designlevel. In this thesis, the control of a washing machine suspension system is achieved byusing adaptive fuzzy sliding mode control methods. The design and implementations are carried out in MATLAB Simulink environment. In the thesis, first, the methods that aregoing to be used and their advantages are given. Then, the unbalanced distribution of loadand its effects on washing machines are explained. Finally, the behavior of the proposed suspension system is examined in Simulink environment.
Design and optimization of a fuzzy-neural hybrid controller structure for a rubbertuator robot using genetic algorithms
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Erdemir, Erdem.; Özkan, Mehmed.
This study presents a combination of soft computing techniques, namely back propagation neural network, fuzzy and genetic algorithms that are used to control the Bridgestone Hybrid Robot Arm (BHRA). The workspace of the BHRA's end effector is divided into small segments and the trajectory independent parameters of all these segments are learned by training small size (only three nodes) neural networks for each segment. The structure of these neural networks is based on the physical model, which is derived from the Language-Euler mechanics of the robot arm. To maintain continuity on the small neural networks, we use a basic fuzzy algorithm whose fuzzy membership function parameters are optimized by genetic algorithm (GA). The proposed technique's performance was compared with only-neural network controller and shown to be more accurate in trajectory control for rubbertuator robots. The main goal of this study is to maintain a better off-line control on the rubbertuators by using only small size (3 nodes and one hidden layer) neural networks and a simple fuzzy algorithm with minimal linguistic variables and minimal number of rules. On the other hand, finding a better off-line control ensures that small learning rates will be sufficient for future on-line training control, and choosing small learning rates will decrease the prospect of divergence and the risk of instability in control.
Design and simulation of a mobile tour-guide robot for museum guidance
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Sarı, Serdar.; Akın, H. Levent.
This thesis aims at building a realistic design of a fully autonomous social mobile robot act as a tour guide to the visitors in museums and that can give information about the museum and the exhibits. It should define all the parts robot need such as locomotion mechanism, sensors, batteries, motors, body, control algorithms and methods required to accomplish its objective. Three dimensional modeling of the robot according to this design, building a virtual museum and dynamic agents, based on a real museum, in a simulation environment for further research and progress are also accomplished. Also, to verify the design, a fuzzy controller based on a layered behavior-based approach, has developed to control the robot. The behaviors used by the robot include going to target point, avoiding obstacle, avoiding human in front of it by speaking and waiting the visitors behind who are following him.
A space subdivision framework for virtual reality and haptic applications
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Diktaş, Engin Deniz.; Şahiner, Ali Vahit.
This thesis presents a framework based on spatial subdivision of triangular mesh objects with octrees for virtual reality and haptic applications. For efficient generation of octree representations a method based on Minkowski sums is developed. The calculation of Minkowski sums is achieved geometrically by direct manipulation of planes in 3D-space. For the evaluation of proximity queries a new data structure called the proximity octree is introduced. A technique for tracking of geometric features of convex objects that are in close proximity to a specific point of interest is developed and tested for a number of applications including haptic interaction.
Energy saving routing protocols in wireless sensor networks
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Çelik, Orhan.; Ersoy, Cem.
Developments in wireless communications and electronics have made designing low-cost sensor networks possible. The sensor networks have many application areas such as health, military, home, agriculture, environmental. Because each sensor has to be low-cost, they have very limited battery and lifetime of the network depends heavily on saving energy. One way of saving energy is designing appropriate routing protocols. In this thesis, we propose some new routing protocols that save more energy and increase the network lifetime comparing to the classical protocols especially in networks with a large number of sensors.
Localization of multiple sound sources in three dimensional environments
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Ünal, Murat Engin.; Gürgen, Fikret.
Localization of sound sources has several applications like teleconferencing, speechrecognition, speaker identification, speech acquisition in an automobile environment,sound capture in reverberant enclosures, large room recording-conferencing, and hearing aid devices. One way of finding the location of a sound source is to utilize the direction ofarrival (DOA) values. This value indicates the angle between two lines, first of whichconnects the mid point of the microphone array and the sound source, and second of whichcarries the microphone array. DOA values are usually estimated by first estimating the time delay of arrival (TDOA) value of the signals received from two microphones and thenconverting TDOA estimates to DOA estimates. TDOA value is estimated using the crosspowerspectrum phase (CSP) coefficients.The existence of multiple sound sources in an environment brings two complications to this problem. First of all, the correlation between the sound sources distorts the TDOAestimation procedure. Second, multiple DOA estimates, which are calculated for multiplemicrophone arrays, have to be matched to the sound sources to find the correct intersection points among multiple ones. Another major complication is the generalization of the environment to the three dimensional space.In this thesis, a new method is proposed, for localizing multiple sound sources inthree dimensional environments. The synchronous addition of CSP coefficients method isutilized for finding the undistorted DOA estimates. Then these estimates are clusteredusing a new design-specific, an inconsistency measure based clustering algorithm. Having found DOA estimate triples for each sound source, the location of sound sources aredetermined by finding the intersection point of three cones formed by three DOA values.This intersection is found by first finding a closed formula, which consists of a single variable, for a locus of the intersection, which is a three dimensional path and then findingthe suitable point on this path. Experiments are done on simulation and real acoustical environments. The results are promising considering the complexities that the algorithm faced, and the number of themicrophones that are available for the experiments, and the diversity of them.
Grey prediction based control of a non-linear liquid level system using PID type fuzzy controller
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Kayacan, Erdal.; Kaynak, Okyay,
This thesis proposes a grey system theory based fuzzy PID controller that has a prediction capability. Although fuzzy control theory and grey system theory have completely different mathematical basics, both deal with uncertain information. In the thesis, a short description of both are given and their performance are compared on a non-linear liquid level control system. The grey model developed is examined under several different conditions and it is shown that the proposed grey fuzzy PID controller has better self-adapting characteristics. The simulation results indicate that the proposed controller has the ability to control the non-linear system accurately with a little amount of overshoot and with no steady-state error. It has, in these respects, better performance than the conventional controllers. The thesis is also intended to serve as a first reading on grey system theory and grey prediction based controllers. The fundamental concepts and mathematical basics of grey system theory are therefore explained in simple terms.
Forecasting stock market return using artificial neural networks
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Doğan, Volkan.; Gürgen, Fikret.; Okay, Nesrin.
Stock market’s return prediction is an important concept in emergent markets like Turkey and Brazil. In this thesis, I used artificial neural networks architecture to model and predict stock markets. Istanbul stock exchange indices National-100 and National-30 are used for domestic market, Brazilian stock exchange index, BVSP, is used as the international market. Input space is divided into clusters with statistical clustering technique Expectation Maximization. ANN’s structure Mixture of Experts is used and local experts are assigned to each cluster. While local experts are learning their region of interest, in parallel, gating experts combine the outputs of them to model overall structure. Besides, future returns are predicted based on patterns obtained from past trainings. In financial time series modeling using ANN, I used past returns in simulations. Since we know two investigated markets are highly volatile and chaotic, the volatility factor is added to the analysis as well. Volatility calculated from RiskMetrics™ [30] is also included in the models to capture different dynamic features of the data. Results of our simulations are evaluated by using previously defined and widely accepted performance measures. Another interesting result is gained during our simulations; two countries of different macroeconomic structures show similarities. Interaction between Turkish and Brazilian stock markets is not surprising; during the financial liberalization of the 1980-1990s, Turkey imported many constitutional laws from Latin American economies especially from Brazil.
Robot parts' rearrangement-sensor uncertainty reduction using particle filters
(Thesis (M.S.)-Bogazici University. Institute for Graduate Studies in Science and Engineering, 2006., 2006.) Bayram, Haluk.; Bozma, H. Işıl.; Ertüzün, Ayşın.
This thesis adresses the parts’ moving problem under noisy sensory information. In this scenario, a 2D workspace contains an actuated robot and a set of unactuated parts. The discrepancy between the robot’s and/or the parts’ real and measured positions may lead to jerky movements or even collisions in the parts’ moving problem we are concerned with. In contrast to previous work, sensory data is no longer assumed to be perfect. Hence the robot needs to approximate state information, taking its highly nonlinear nature into account. It accomplishes this using particle filters, which implement a recursive Bayesian filter in nonlinear and/or nongaussian environments. For the model of parts which turns out to be linear, the approach reduces to Kalman filtering. First the robot’s dynamic model and the measurement model are modified to incorporate the inaccuracies in the sensory data; and then the particle filter is utilized to get improved positional estimates. Enhancements in the robot’s movements and reduction in the number of collisions have been verified through extensive computer simulations. An evaluation of its theoretical performance is presented based on the Cramer-Rao lower bound. Finally, a series of experiments with EDAR provide insight into real-time performance.

Browse

Browsing M.S. Theses by Issue Date

Results Per Page

Sort Options