PI: V. M. IstemihanGenc

Researchers:Yusuf Yaslan, Yaprak Yalcin, Dr. Mohammad Mahdi

TUBITAK 1001 – Scientific and Technological Research Projects Funding Program, 2018-2020, Project no: 118E184

Catastrophic faults that lead to transient instability in electrical power systems may trigger fast evolving widespread outages such as blackouts. In recent years, Wide Area Measurement Protection and Control (WAMPAC) systems are adopted in modern power systems to increase the system observability and security. In this project, it is aimed to detect the instantaneous and fast propagating transient instabilities that are occurring in large scale interconnected power systems faster (earlier) than real time, and to prevent system collapses by applying emergency corrective control measures within the WAMPAC environment.In the project,firstly, a transient stability prediction method based on multi-channel convolutional neural networks using time series of PMU measurements and wide-area measurement-based transient stability prediction using long short-term memory networks are developed. Afterwards, an emergency control scheme is developed based on decentralized model predictive control (MPC)to prevent transient instabilities occurring in power systems.  After an onset of transient instability due to a contingency, the control scheme acts on the intercept valves of the steam turbines and the excitation systemsof the generators that are available for stabilizing the system.  Using the proposed method,  thesystem operational constraints as well as the turbine actuator dynamics are properly taken into account, and an objective function towards stabilization is minimized at every sampling time to generate an optimal control law.  The method encompasses a decentralized control strategy allowing a straightforward implementationthat  does  not  require  a  communication  system,  nor  does  it  suffer  from  its  inherent  delays,  where  each participating generator unit is equipped with an MPC controller that relies on only local measurements.Following a  critical contingency,  the controllers  force the  generators to remain  in synchronism  and bringthe  system  back  to  a  stable  operating  condition.   The  performance  and  efficacy  of  the  proposed  methodhave been evaluated and demonstrated by simulations performed on two test systems, the WSCC 9-bus andthe IEEE 68-bus systems.  It has been shown that the controllers, which are designed independently fromthe contingencies and operating conditions, are effective in restoring the system’s stability and are robusteven when the system is subjected to structural changes due to integration of power electronics interfacednon-synchronous generator units

PI: YaprakYalçın,

Researchers:AnılYılmaz

ITU BAP Research Grant, 2019-2021, Project code:MYL-2019-42167

In the project, consensus (common task fulfillment) control problem in autonomous vehicles is considered.  An MPC control structure is developed for the solution of the problem taking into account uncertainties and changes in vehicle parameters and communication speed. In this project, after linear MPC is mainly designed for consensus control, parameter estimation with the "Immersion Invariance" method is established in the discrete-time setting (to obtain vehicles with different and variable masses by placing different weights on the vehicles). Finally, MPC is combined with this constructed estimator and an adaptive MPC consensus is established.

PI: Volkan Sezer,

Tubitak 1001 Project, Starting in March 15, 2022

The research and development activities we plan to do in this project are summarized in 4 main sections as follows.The first problem is the process of estimating the position of the robots whose initial position is not certain (global localization). Solving the global localization process quickly and accurately is an important research topic in autonomy. The second problem is the correct tracking of the robot's position on the map (local localization/position tracking), whose initial position is known. In most of the solutions for this problem, markers on the map are used. In the proposed project, we will work on new position tracking solutions to provide more accurate results under real-world constraints such as noisy measurements in dynamic environments and inaccurate mapping, by obtaining virtual and dynamic markers on the map. The third problem is the "local planning" problem, which is one of the most critical components of autonomy. In the project, the local planner will be based on the Follow the Gap Method (FGM) method. In the proposed project, an intelligent local planner will be developed to eliminate various disadvantages. In addition, in the case of moving obstacles, new solutions will be offered by making the selection of spaces in a predictive way.The fourth subject is the "human following feature" that some of the autonomous robots have. Many robots working in the factory, used for military purposes or in the service industry must have. FGM's ability to safely avoid obstacles while heading towards a local target and that it has not been applied to a goal following problem before, will provide a unique solution to this problem. In addition, special situations such as the movement of the person being followed, the human’s state estimation or disappearance of the human in the field of view will be studied forthe solution of human following.

PI: Volkan Sezer,

Tubitak 1001 Project, 2019-2021

Within the scope of the project “118E809”, new methods that can be used in autonomy and other fields have been developed, together with an autonomous wheelchair that can operate fully autonomously in closed environments. First of all, the hardware design of the chair was realized and it was transformed into “drive by wire”, together with its sensors and processing units. The chair can obtain the map of the environment by utilizing "simultaneous localization and mapping" techniques, and then it can continuously calculate its position on this map. In addition, it can plan its path to the destination and safely avoid external obstacles while tracking the path. In order to perform collision check while calculating the global plan; a new method has been developed that can test whether a point is within the polygon very quickly according to the Voronoi regions. From this point of view, a new classification method that makes use of the Voronoi regions was also developed during the project studies. In addition to these, a new and safe local planner has been developed which can pass farther than obstacles, based on the "follow the gap" method, The developed autonomous platform can plan its velocity while tracking the path by considering environmental factors. A new fuzzy logic based velocity planner that takes into account both mass and comfort for the first time in literature, has been developed. All these developed methods have been tested in simulation environment and experimental studies. Comparative analyzes of the methods are revealed and the benefits of the methods are demonstrated. The studies have been published as international journal articles, book chapters, conference proceedings, graduate/undergraduate thesis, and there are studies that are accepted or still in the review process. The developed autonomous wheelchair platform is planned to be used for the future autonomy solutions.

PI: Volkan Sezer,

Researcher: Tufan Kumbasar,

ITU Research Grant, 2018-2019

As a result of rapid developments in sensor, actuator and computing units, robotic systems are fastly becoming a part of daily life . Service robotics is defined as an important sub-topic in robotic systems. These types of autonomous robots are very popular in indoor transportation in the Industry 4.0 concept and their usage is increasing day by day. Autonomous ground robots (AGV), which started to be used in the factory together with Industry 4.0, now move mostly on fixed routes determined for them and take their position directly from the environment. In the future, these robots are expected to be more intelligent and make more autonomous decisions.Within the scope of this project, various improvements have been realized in order to make autonomous robots to work in indoor areas more efficient. The improvements have been realized in the areas of Simultaneous Positioning and Mapping (SLAM), one of the most critical problems of autonomous robots and planning / control. In order to solve the SLAM problem faster, IDA-SLAM approach has been developed which enables the robot to look at a narrower area more efficiently, not all the existing pointers during the data association phase. Faster solutions are obtained with this approach. In the planning and control section, a reinforcement learning approach has been developed which enables the robots to plan by considering the loads they carry on them. The efficiency of these two studies compared to the standard approaches in the simulation environment and the real-time operation of the methods was confirmed by tests on real robots. In addition, fuzzy logic and autonomous robot control have been studied in this project.

PI: Mehmet Turan Söylemez,

Co-PI: I Ustoğlu, A Deniz

Researchers: S. Boynukalın, Z Öztürk,SKurutulan,  M Yeşioğlu, M Başkak, C A Bayraktar,  T Yüksel

ITU Arı Technopolis, 2021-Today

A holistic and inclusive signaling action plan for TCDD railway network signaling systems for purposes such as improving operational safety on TCDD railway lines, reducing travel times, increasing traffic safety, using existing road capacities effectively and efficiently, increasing mobility, using local resources and opportunities in signaling more effectively will be done.

In this context, the Railway Signalization Action Plan, which takes into account the new developments in the country and in the world, will be prepared with the participation of wide stakeholders in order to be a roadmap for the studies to be carried out on railway signaling in the short, medium and long term, and the necessary strategies and methods will be revealed.

PI: Mehmet Turan Söylemez,

Co-PI: E Dincel

Researchers: M.F. Özcan, S. Boynukalın, Ç.C. Erdem, O Bakırcı, M Canevi

ITU Arı Technopolis, 2020-Today

In modern rail systems, train operation is carried out by automatic signaling systems, not by drivers. Within the scope of this project, “Automatic Train Operation Speed Regulation Algorithm Design”, which is a part of the signaling system that ensures safe, punctual, and energy-efficient train operation, will be carried out. As a result of the project, the developed automatic speed regulation algorithm will enable the trains on a subway line to be driven by the CBTC signaling system considering different driving profiles and the P.I.D. required for train control. The design of the controller will be provided by the research and development of other similar systems in the world. The algorithms to be developed within the scope of the project and the speed profiles depending on the use of different performance parameters that the operator can use depending on his request will be poured into the software to be developed and the design will be carried out. With the controller, it will be stopped at the stopping point with a precision (a few centimeters) determined at the stations. Apart from speed profiles depending on performance parameters, eco-driving mode will also be developed within the scope of the project and included in the algorithms. Algorithms developed in the project will be tested on the trains supplied by the CRRC company on the Gayrettepe-Istanbul Airport Metro Line being built in Istanbul and will be delivered to ASELSAN at the end of the project.

PI: Mehmet Turan Söylemez,

Co-PI: S. Açıkbaş

Researchers: S. Boynukalın, Ç.C. Erdem, O Bakırcı

ITU Arı Technopolis, 2021-Today

Railway signaling systems have been developed and started to be used in order to ensure that the train traffic on the railway lines is carried out safely and to use the railway lines more efficiently. In parallel with the developing technology, improvements have been made in signaling systems, and today, signaling systems with train control systems that fully control the train have been produced. The ERTMS train control system is a combination of ETCS and GSM-R used in railway communication, and it is an important signalingsystem that has been implemented to eliminate the system differences experienced in the signaling systems of European countries, both within and between countries, in recent years. Within the scope of this project, the ERTMS/ETCS On-Board system will be developed by ASELSAN. In this context, there is a need to develop the braking curve below for the relevant system in accordance with the requirements specified in the scope of work. The ERTMS/ETCS onboard system will be used in level 1 and level 2 systems. Note: Related documents can be obtained from https://www.era.europa.eu/content/set-specifications-3-etcs-b3-r2-gsm-r-b1_en.

PI: Mehmet Turan Söylemez,

Co-PI: S. Açıkbaş

Researchers: U. Yıldırım, M.F. Özcan, A Polat, S. Boynukalın

ITU Arı Technopolis, 2016-2020

Within the scope of this project, a rail traction power system simulation server software and a rail system simulation client software have been developed. During the development of this software, the V&V model, which is the standard in the development of such software, has been followed by using the waterfall workflow method. In this context, first of all, software requirements were revealed, high-level designs have been made in line with these requirements, the necessary program modules have been then determined, and detailed designs for each module were passed to the coding phase. After the unit tests for each module, the integration of the program has been ensured and the system level testing and validation processes and program development have been finalized.

PI: Mehmet Turan Söylemez,

Co-PI: S. Açıkbaş

Researchers: M.F. Özcan, S. Boynukalın

ITU Arı Technopolis, 2020-2021

Within the scope of this project, it is aimed to carry out detailed traction power system (electrification) simulation of modern rail systems as well as operational simulation (in a way that can optimize fixed and moving block signaling) through licensable software, where the user can follow the movements of the trains. As a result of the simulation, an optimized time-table/time-table output can be obtained. In addition, an environment has been provided where the user can compare the previous simulations and easily display their data graphically. With its flexibility in speed profile and headway calculations, the ability to perform traction power (electrification) and motion simulation of urban and intercity railway enterprises has been gained. In this way, the system modeled during the simulation can be monitored and the outputs for the optimum use of the capacity of the lines can be provided with the reports. Within the scope of the project, it is aimed to add new features to the software developed in accordance with the client-server architecture, to model detailed train motion simulation on the client-side, and to add the ability to make more detailed analyzes of the train motion simulation available on the server-side.

PI: Tufan Kumbasar,

Researchers: Aykut Beke, Serkan Türkeli

İTÜ BAP Research Grant, 2017-2018

In this project, the frameworks Type-2 Fuzzy Logic, Control Theory and Image Processing have been used to provide an intelligent pursuing strategy generation for the area of pursuit-evasion games. This research project presents a systematic and interpretable design approach to generate Type-2 (T2) fuzzy logic based linguistic Pursuing Strategies (PSs) and their deployment to a real-world pursuit-evasion game named as “Game of Spheros”. Firstly, we have developed a novel T2 Fuzzy logic based Strategy Planner (T2-FSP). Then, through detailed theoretical investigations on the input-output mapping of the T2-FSP, it has been shown that it is possible to design a linguistic PS which defines both pursuer’s approaching behaviour (Aggressive, Smooth) and side (Left or Right) to the evader by simply tuning the Footprint of Uncertainty (FOU) sizes of the T2 fuzzy sets. Hence, an interpretable relationship has been revealed between the FOU sizes and the PSs through comparative theoretical explorations and derivations. A real-world game environment is constructed in order to validate the developed T2 fuzzy logic based PSs in real-time. The game scenario is constructed as the evader droid BB8 will be controlled by a human user while the pursuer droid Sphero 2.0 will be navigated through the game environment via the proposed T2 fuzzy pursuing system. The proposed T2 fuzzy pursuing system structure is composed of vision based localization, the error signal generator, T2 fuzzy strategy planner and the control system. Comparative experimental results have been presented to show that the T2 fuzzy logic based PSs have a satisfactory performance against a human user.

PI: Tufan Kumbasar,

Researchers: İlker Üstoğlu, Serhat Şeker

TÜBITAK 1001 Research Grant, 2019-2022

In this project, Deep Learning-based T2-FLS (dT2-FLS) layers are developed for learning multi-input multi-output nonlinear dynamic and static systems. The performances of the structures are examined with the state of the art methods through data collected from both simulation and experimental environments. For the realization of this objective, it is evident that there is a need for interdisciplinary academic study. This project took up fuzzy logic, system theory, and deep learning frameworks and investigated ways how to employ these frameworks in the design of dT2-FLS to provide a novel contribution to the literature.In this project, the structures of T2-FLSs are examined to extract their mathematical expressions. Then, these are directly used to accomplish the training of T2-FLSs via deep learning training methods. The proposed dT2-FLS layers are also integrated into multi-layered hierarchical (parallel or serial) neural network architecture. The proposed structures are also blended with deep learning layers to develop hybrid structures.

Project manager: Serhat İkizoğlu

Consultant: Ahmet Ataş

Researchers:KaanŞahin, Eyüp Kara, Saddam Heydarov, TunayÇakar

Granted by: TUBITAK (1001)

Project code: 115E258

Starting date: April, 2015

Completion date: May, 2018

This is a multi-disciplinary project bringing researchers together from Control and Automation Engineering Dept. (Istanbul Technical University) and Audiology Dept. (Istanbul University – Cerrahpasa Medical School). The project is about determining the problems of individuals suffering from balance disorder. The originality in this work lies in the fact that –as the first time in the literature- we make use of data collected synchronously both from the insole pressure sensors and the motion sensors placed on the body. A two-step procedure is considered for the overall method. The first step serves for a binary classification where the subject is decided to be a member of either the healthy group or the patients’ one. In case the decision is ‘the second group’, a second step follows that specifies the problem. Since initially a high number of features (33 major – totally 150) is determined, a dimensionality reduction is applied as the next step in order to mitigate the algorithm complexity and to prevent the machine from overlearning.

The target is reached for two diseases. The overall success is calculated as 93.8%. Features based on force and motion sensor data and those related to correlations between them contributed in this success which was a nice proof of our prediction for the originality of the project.

The findings give high motivation to continue with the study. Though the project duration is over, we go on working on the matter. To increase the performance of the system, on one hand we try to increase the data volume while we search for new features on the other.

So far, the studies on the project have produced 4 journal papers and several conference papers.In addition, many master's theses have been completed in the relevant field.

PI: Müjde GÜZELKAYA,

Researcher: Erhan YUMUK

ITU BAP - Doctoral Research Grant, 2018-2022

In the doctoral research, it has been aimed to design a controller for systems represented by fractional order models. In this respect, simple fractional order models with or without time delay are selected and four different studies on controller design using these types of fractional order models are carried out. The first study provides a novel optimal fractional order design methodology using direct synthesis method for single fractional order pole model without time delay. The second study offers an analytical fractional order controller for simple fractional order models with time delay. By this study, the notion of delayed Bode’s ideal loop transfer function is introduced to literature. In the third study, it has been shown that this methodology is suitable for real-time applications using magnetic levitation system. The fourth study proposes a robust fractional order controller design based on delayed Bode’s ideal loop transfer function. In these four studies, closed loop configuration is employed. On the other hand, the last study focuses on tuning of modified posicast controller which is an open-loop control method.

The Turkish Accelerator Center (TAC) Project is proposed as a national center that will contain MeV and GeV scale electron and proton accelerators. The Turkish Accelerator and Radiation Laboratory in Ankara (T.A.R.L.A) is an IR FEL oscillator and Bremsstrahlung facility and planned as a first facility of the TAC project. The building of TARLA facility has been constructed in Gölbaşı Campus area of Ankara University in 2011. It is proposed that TARLA will be operated by 20 kW solid state power amplifiers. The RF power transmitted into the TESLA RF cavities will be net 16 kW. Since there is no available 1.3 GHz 20 kW solid state power amplifier, it will be a new development for TARLA .low noise and stability are indispensable requirements in this application. For this reason, a compact and a robust L band 20 kW RF amplifier design has been planned by using multiple gallium nitride (GaN) RF high electron mobility transistors (HEMTs) due to high power capability, stability and noise immunity at a given frequency range. Multi-stage Class AB structure will be utilized for inherent linearity and efficiency properties.

In this project work covering the years 2013-2017, studies were carried out under the titles of Bunch Dynamics and Bunch-Beam Interactions for the CLIC collider. The selection, design and optimization of the basic equipment (source, injector, RF cavities, magnets, etc.) that will shape the beam dynamics for the targeted beam quality in the design of particle accelerators is of great importance. Beam-beam interactions, on the other hand, are related to the design of magneto-optical systems, beam optimization, monitoring of background formation, etc., especially around the collision point. It deals with the elaboration and analysis of issues. Various studies were carried out on the topics mentioned within the scope of the project. During the project period, studies were also carried out on the design of the XFEL facility based on CLIC technology, and a partnership was made with the H2020 project named COMPACTLIGHT, which focuses on the technical design of the X-Band X-FEL facility.

The visits of our researchers to CERN within the scope of the project were supported by TAEK.

PI:

Researchers: Ozlem Karslı , Mustafa Dogan, FatihAhiska, and O. OrkunSurel

TUBITAK 1001 Project  2015-2017

Project associated with Turkish Accelerator and Radiation Laboratories at Ankara. Grant no: 114E085 L Band High Power RF Pulse Compressor, Institute of Accelerator Technologies, IAT-TAC, budget: 190,000 USD, Senior Researcher.

High-frequency power amplifiers can be equipped with complementary pulse-compressor systems to fulfill the

advanced cavity power requirements. Microwave pulse compression is a power extraction method based on pulse width reduction and electronic triggering. The cavity is fed by pulsed-RF source, the resonance is evolved at t ≥t0, t0 = cavity filling time. Established standing waves or resonance condition is broken down by triggered plasma switch programmed with adjustable delay and pulse width. This causes the stored energy to be transmitted as shorter output RF-pulse. The implementation of a microwave pulse compression system requires careful measurements and structural modifications to be able to tune design parameters for optimal setup. In the scope of this paper, the experimental results are obtained for different waveguide lengths and various load configurations (to exploit new opportunities). All connections, iris configuration, and precise adjustment of plasma triggering have been set up cautiously. The cavity gain is increased up to 23 dB with trigatron driving improvements. High power requirements and demands are investigated for many applications, particularly for biomedical science projects.

Tubitak Project (2017-2021)

Temsa – İTÜNOVA

ITU Side: Executive: Prof. Dr. Metin Gokasan

Researchers: Prof. Dr. Seta Boghosyan, Asst. Prof. Dr. Tankut Akgul, Lect. Ali Bahadir

In the project, the hardware and software of the power supplies that will generate variable amplitude and frequency AC power sources to control all the electrical actuators in the system from the high voltage DC voltage source, and variable average value DC power supplies for battery charging, are designed, implemented and used in the vehicle with minimum size and high efficiency in a box. will be realized. In the vehicle propulsion system, the hardware and software design of a high-efficiency driver that can control the PMS engine in 4 regions will be realized. The packaging of the battery system, which will be selected considering the vehicle dynamic parameters and the requirements expected from the vehicle, will be carried out by using the previous experiences of the company. High efficiency SiC-based power switches will be located on the power floor of the motor driver, and a controller developed for functional safety applications in accordance with ISO 26262 and IEC 61508 standards will be used as the controller. In the control algorithm, there will be EKF or SMO estimation algorithms of the state and parameter values, with the obtaining of the reference currents that will create the maximum torque per current, and the durable control algorithms with sliding mode in the current control. SiC and GaN elements will also be used in DC-DC converter, which is another high efficiency power stage to be developed. The entire system will be placed in a box and will be optimized in terms of both size and cooling systems.

1007 Program for Supporting Public Institutions Research and Development Projects

2016-2022

CUSTOMER INSTITUTION: Ministry of Transport, Maritime Affairs and Communications –

T.C. General Directorate of State Railways Administration (TCDD)

Managing Organization: TÜBİTAK MAM - ENERGY INSTITUTE

Executive Organization: SAVRONİK Elektronik San. and Tic. Inc.

Executive Organization: SONMEZ TRANSFORMATOR SAN. VE TİC. Inc.

Executive Institution: ISTANBUL TECHNICAL UNIVERSITY

Executive Organization: MEDEL ENGINEERING AND ELEKTRONİK SAN. VE TİC. Inc.

Executive Organization: TÜLOMSAŞ GENERAL DIRECTORATE

ITU Team:

Prof. Dr. Metin Gökaşan (Executive Assistant Conv.)

Prof. Dr. Ata Muğan (Bogie Carriage An. Test, Vehicle Body Carrier An. Test)

Prof. Dr. Seta Boghosyan (CerConv.)

Assoc. Prof. Dr. Güven Kırış (Serve Transf.)

Assoc. Prof. Dr. Volkan Sezer (CerConv.)

Assoc. Prof. Dr. Tufan Kumbasar (TKYS)

Prof. Dr. M. Turan Söylemez (Emn. Fnk)

Asst. Prof. Dr. İlker Üstünoğlu (Emn. Fnk)

Asst. Prof. Dr. Özgür Türay Kaymakçı (Emn. Fnk)

Asst. Prof. Dr. Murat Yılmaz (Traction Motor)

Asst. Prof. Dr. Çiğdem Kaya (End. Tas.)

Assoc. Prof. Dr. Emin Sünbüloğlu (Ent. and Static Tests, Loko. Fnk. and Acceptance Tests)

Asst. Prof. Dr. Vedat Temiz (Ent. and StatikTestler, Loko. Fnk. and Acceptance Tests)

Dr. S. Ergün Bozbağ (Ent. and Static Tests, Loco. Fnk. and Acceptance Tests)

The electric E5000 Mainline Locomotive, which will be designed and produced in this project, will be a locomotive that can be sold in the country and abroad by completing its homologation with TSI certification. In addition, all the main equipment developed are products containing important and critical technology that can be sold as separate products to the country and abroad. The products in question will be able to be used in the existing locomotives within the body of TCDD, both as spare parts and within the scope of their modernization. ITU will take part in this project as teams in all work packages. He will take part in the design of the vehicle's body, bogie, traction system, auxiliary power systems and the development of prototype tests, as well as taking on very important tasks in certification and acceptance tests.

2012K120730, Ministry of Development, 2012-2021.

Coordinator: Prof. Dr. Metin Gökaşan

Researchers: Prof. Dr. Seta Estrada, Assoc. Prof. Dr. Zeki Y. Bayraktaroğlu, Assoc. Prof. Dr. Volkan Sezer

The total vehicle production in the world in 2010 was around 60 million, and it is estimated that the number of vehicles active on the highways around the world exceeds 600 million. [International Organization of Motor Vehicle Manufacturers, http://www.oica.net/] Considering the situation in Turkey, 870 thousand vehicles were produced in 2009 and it is known that the number of active vehicles on our highways is over 14 million currently. Our country is among the 16th European Union countries in world automotive production; it ranks first in light commercial vehicle production, second in bus production, ninth in heavy truck production and seventh in total. Considering the place of Turkey in vehicle productionin the world, land vehicle technologies emerge as a dominant industrial field. Considering the environment-friendly, sustainable development and growth issues, there are three main areas where this great production and activity created by the tire road vehicle is in a deadlock:

(1) The energy resources are based on a single exhaustible area, fossil fuels,

(2) Harmful effects on the environment and human health arising from the use of existing vehicle technologies and road infrastructure: air pollution, noise pollution, traffic density,

(3) The increase in traffic accidents, which are directly related to the increasing number of vehicles and mainly caused by human error, and their burden on the society and the national budget. Within the scope of this project, as a result of the research and development activities to be carried out by taking Hybrid and Electric Vehicles (HEV) as a development platform, not only the problems (1) and (2) as usual, but also the integration of vehicle active safety systems based on the developments made on the HEV and smart traffic networks. At the same time, problem (3) is also targeted.

Tubitak1003 Project115M092 (Advisor: Metin Gökaşan), 2016-2018.

Manager: Prof. Dr. Ata Muğan

Diesel engines have been produced in Turkey and in the countries to which we export products. On the other hand, the Electronic Control Units (ECU) in all these products are supplied from abroad by importing from a number of foreign companies and dependence on these companies emerge.This dependence on ECU producing companies occurs not only in procurement of ECUs, but also in development of ECU software and in future modifications. The main aim for initiating this project is to eliminate this dependence and to produce the crucial component “diesel engine control system” by the use of domestic sources and software. By the use of proposed ECU system, actuators of internal combustion engines will be controlled, variables of engine will be measured and unmeasureable variables that will be used in control algorithms will be calculated by observers.The internal combustion engines have been produced in Turkey through the license agreements and contributions of some domestic companies. On the other hand, to catch the continuously tightened emission standards cannot be achieved by just producing a good quality engine. Besides, control of the engine and development of its control algorithm are also very important.The engines have nonlinear dynamics and its characteristics need multidisciplinary engineering approaches. The studies on this subject depends on realization of model based control algorithms to achieve desired targets and control of many actuators in real-time to achieve these targets. Moreover, all variables could not be obtained by direct measurements. Dynamicsof variables are different from each other and some of them have high frequency signal contents. As a result, the microcontroller hardware on which the control algorithms will be run has a complex nature. The engine model inside the ECU will have the following submodels: EGR, VGT, throttle, common rail, average combustion, mechanical loading and traction control. It will also have interface with the peripherals of the vehicle control system. The ECU that will be developed through this project will be utilizing the complex optimization algorithms to generate the reference values produced by the control algorithm while it is controlling EGR, VGT, throttle, common rail pressure, fuel injection timing and durations.

PI: Prof. Dr. Gökaşan,

Researcher: Assoc. Prof. Dr. Tufan Kumbasar

The control algorithm design, which is one of the critical components in the DIKONS Project, will be realized. According to the International Maritime Organization (IMO) and certification class societies (DNV, ABS, LR, etc.), a dynamically positioned (Dynamic Positioning-DP) ship is able to maintain its position and direction (fixed groundor predetermined orbit) through its active thrusters against wind, wave, tide and current is defined as a ship with a system infrastructure. The Dynamic Positioning (DP) System is also a system used when the ship is not able to anchor or wants to anchor, or when it needs to stay in a sensitive position. A dynamically positioned ship uses existing propulsion systems to stay on a fixed course or position, and the ship's propeller, rudder, bow and stern thrusters and all other combined systems work together against the wind and current to keep the ship's bow in the desired position. For this purpose, the location information of the ship; it is used by the DP control system with information from DGPS, Gyro, wind and motion sensors.

TubitakARÇELİK1505 Project 5130069,2014-2017

(Coordinator: Metin Gökaşan)

Researchers: Prof. Dr. Fikret Caliskan, Assoc. Prof. Dr. Ali Fuat Ergenc, Asst. Prof. Dr. Murat Yilmaz

Arçelik A.Ş. company uses BLDC/PMSM motors in most of their current white goods products. With the field-oriented control (FOC-Field Oriented Control) of the BLDC/PMSM motor used in the drum drive of the washing machine, an increase in system efficiency in washing/spinning, a reduction in energy consumption and an improvement in motor-induced sound quality are achieved. In case the FOC method is performed without a sensor and with an affordable microprocessor, in addition to the above gains, Hall sensor assembly, workmanship, possible SAO problems and cost reduction are targeted. The FOC methods used in Arçelik's current products are the methods developed by foreign microprocessor companies and used in their products. The use of these methods not only solves the daily need for Arçelik products, but also makes Arçelik dependent on foreign microprocessor companies. This resulting dependency creates a disadvantage in terms of time, financial resources and flexibility in case of dissemination and adaptation of the same method to a different product, and adversely affects Arçelik in competition.

PI: İlhanKocaarslan

Researchers:

TUBITAK, 2020-2021

In the “Blockchain-based Renewable Energy Resource Guarantee (YEK-G) System” project, under the coordination of the Energy Markets Operations Inc. and the Energy Market Regulatory Authority, the production of renewable energy plants is defined in a blockchain-based system with a digital green certificate, and the development of all energy market stakeholders throughout Turkey. The YEK-G system has been developed, which allows them to trade green certificates at regular intervals over a digital platform, end-to-end, reliable, verifiable and traceable to the past. While similar systems in other countries do not include blockchain technology, in Turkey, with this cooperation, blockchain is a transparent, reliable system, historical records cannot be changed, can be traced retrospectively and its autonomous features have been successfully applied to the renewable energy certificate market concept. The developed system is open to the participation of all energy market stakeholders in Turkey and it is served since June 2021.

PI: İlhan Kocaarslan

Researchers:

TUBITAK, 2017-2021

In this project, which was accepted within the scope of TUBITAK ARDEB 1003, 216M252, a hybrid technology that can be used in internal combustion engine vehicles, increasing fuel efficiency and accordingly reducing carbon dioxide emissions will be developed. Combining the internal combustion engine, BLDC engine and Thermoelectric Generator technology to increase fuel efficiency by around 15%, to increase the competitiveness of our country with the production of high value-added technological products, to train expert scientists in this field, to publish qualified publications within the scope of SCI, and to obtain patents if the project is successful. It is aimed to develop this system in large-scale commercial vehicles with high fuel consumption (such as trucks and buses) by optimizing the system in the light of the results to be obtained.

PI: İlhan Kocaarslan

Researchers:

TUBITAK, 2018-2020

In this project, which was accepted within the scope of TUBITAK 1501-OTO-HEAT 2015-2, it is aimed to design and develop a motor and motor driver, which does not use rare elements such as magnets, to be used in electric vehicles, to make them ready for mass production and to use them in an application tool, which is also determined as the project output.

PI: İlhan Kocaarslan

Researchers:

TÜVASAŞ, 2019-2020

N85 Personnel Accommodation Vehicles, designed in 2019 for the purpose of accommodation of TCDD personnel during road works, consist of 4 bedrooms and a living room with an American kitchen. These wagons, produced as 40 in total, are also offered to the public in order to meet the need for shelter in case of natural disasters such as earthquakes.

PI: İlhan Kocaarslan

Researchers:

TUBITAK, 2019-2021

This project was called by TÜBİTAK in order to manufacture composite brake pads, from which all domestic needs are imported, and the project was given to TÜVASAŞas a result of the evaluation. It is possible to say that the amount of brake shoes used in railway vehicles and the imports made accordingly will increase at a very high rate. If it is not available, it will bring the rail transport to a standstill.

PI: İlhanKocaarslan

Researchers:

TÜVASAŞ, 2017-2020

In order for modern railway vehicles to have an assembly line with today's technologies, a railway vehicle interior lining assembly line was established. High-speed trains, city rail system vehicles and conventional railway vehicles will be assembled on this line.

PI: İlhan Kocaarslan

Researchers:

TÜVASAŞ, 2017-2020

Bogie chassis robotic welding production facility was established due to reasons such as the strategic importance of the bogie chassis, which is one of the most important parts of the railway vehicle, being one of the safety elements, and the high need for bogie chassis in the domestic and international markets. The production facility has two robotic cells and three large positioners, and 96% of the welded production is carried out robotically.

PI: İlhan Kocaarslan

Researchers:

TÜVASAŞ, 2019-2020

It is aimed to be able to produce imported high-speed train sets with domestic and national means. In this context, it is planned to increase the speed of the current 160 km/h National Electric Train Set to 225 km/h. A configuration was created so that each high-speed train set consists of 8 motor vehicles.

PI: İlhanKocaarslan

Researchers:

TCDD Railway Research and Technology Center, 2017-2020

In the project, it is aimed to produce the imported train sets with domestic and national resources. It was aimed to meet the electric train set needs of TCDD first and then to contribute to the country's economy by exporting the train sets. The first prototype train sets produced are the first domestic train sets of the Republic of Turkey, each trainset consists of 4 driven vehicles and has a speed of 160 km/h.

PI: İlhan Kocaarslan

Researchers:

AB Horizon 2020, 2017-2020

The Cities for People project, accepted within the scope of the European Union Commission's Horizon 2020 "Mobility for Growth" call, is a 3-year Research and Innovation project. Within the scope of the project, citizens, city officials and innovation experts in the five developed and developing city regions (Üsküdar-Turkey, Hamburg-Germany, Budapest-Hungary, Oxfordshire-United Kingdom, Trikala-Greece) will be able to understand the transportation and mobility challenges and priorities, to develop ideas and concepts and will work together as a “community” to produce, test these concepts in real life and develop their potential.

PI: İlhanKocaarslan

Researchers:

Istanbul University Research Grant (BAP), 2015-2018

In this project, it is aimed to design, model and simulate the Switchable Magnetic Field Brushless DC Motor and Driver with MATLAB and PSIM programs and to use the final prototype in electric vehicles.

PI: İlhanKocaarslan

Researchers:

Istanbul University Research Grant (BAP), 2016-2017

Modeling, simulation and realization of Unmanned Aerial Vehicle Design and Implementation.

PI: HakanTemeltas,

Researchers: Abdullah Ersan Oğuz

Scholars: Mehmet Caner Akay, Osman Ervan, Abdulbaki Aybakan

TÜBİTAK 1001 Research Grant, 2017-2019, Grant No: 116E178

Simultaneous implementation of SLAM method and the motion planning for robot teams is called Active-SLAM problem. The purpose of this concurrent realization is to realize the planned routes for autonomous robots while minimizing the uncertainty in both localization and mapping. The main purpose of this work is to design an Active-SLAM system based on informative motion planning for heterogeneous robot teams. Relative entropy theory will be presented for the information maximization approach. By this way, the heterogonous team plans their paths so that total uncertainty always decreases. The methodology developed under this project will be tested experimentally on the heterogeneous robot team with laser distance measurement systems (LIDAR). Several conference papers and journal articles are published within the scope of this project (1 SCI-E journal articles, 4 international conference papers). Moreover, one PhD and two master theses are supported within the scope of this project.

PI: Hakan Temeltas,

Researchers: Serhat İkizoğlu

Scholars: Abdurrahman Yılmaz, Ecem Sümer, İbrahim Dinçer Nalbant

TÜBİTAK1003 Research Grant, 2018-2020, Grant No: 116E734

The automated guided vehicles (AGVs) are used in factories to perform logistics andmanipulation tasks. While performing these tasks, for example moving from one point to

another, they can only perform pre-programmed tasks or they need to keep a predeterminedpath. The aim in Industry 4.0 compliant smart factories is that the smart AGVs to be used insmart factories are self-organized and capable of using entire free space in the factory. In orderto develop such a fully autonomous structure, a method using the two-stage localizationapproach has been proposed within the scope of this project. Self-Adaptive Monte CarloLocalization (SA-MCL) approach has been used in large-scale localization to determine theinitial poses of AGVs if unknown, and to track their positions in the regions far from the target.The Ellipse Based Energy Model has been proposed to increase the global localizationperformance of the method. On the other hand, to estimate the poses of the AGVs with highaccuracy and dock to the target with sub-centimeter error, Affine Iterative Closest Point (AffineICP) method has been utilized for fine localization in the regions close to the target. In order toswitch between large-scale localization and fine localization approaches, a hysteresis curve-based decision mechanism has been developed. The methods and approached proposed inthis project are validated with simulation studies and experimental field tests. Several conference papers and journal articles are published within the scope of this project (4 SCI-E journal articles, 4 international conference papers). Moreover, one PhD and two master theses are supported within the scope of this project.

PI: Fikret Çalışkan

Project Type:General Research Project

While detecting and isolating the sensor and actuator faults that can occur in unmanned aerial vehicles (UAV), which are frequently encountered in the literature, either a linear mathematical model of the vehicle or a first-order linearized model around an operating point is taken into consideration. The standard Kalman filter (KF) is commonly used in linear models, and the Extended Kalman filter (EKF) is used in first-order linearized models.

However, real aircraft dynamic models are high order nonlinear models and the standard Kalman filter cannot be applied to this nonlinear model. The extended Kalman filter gives an approximate first-order solution. The estimation errors increase as you move away from the operating points where linearization is made. In this project, it is proposed to produce a solution for vehicle models(UAV) with high degree of nonlinearity, Unscented Kalman Filter (UKF) and Cubature Kalman Filter (CKF), which can predict with higher accuracy, will be used for non-linear aircraft model. Thus, fault detection and isolation can be performed more safely. The results obtained will be compared with the results obtained with the Extended Kalman filter and with each other. The results will be tested on a real quadrotor.

PI: Fikret Çalışkan

Project Type:Thesis Project, Ph.D.

The subject to be focused on as a result of the thesis project to be carried out; the navigation of a four-propeller rotor system in an interior space, thereby automatically performing visual inspection of aircraft surfaces while flying autonomously. Artificial intelligence algorithms will be used while performing the visual inspection operation on aircraft surfaces. For indoor navigation, by using the beacon technology and the supplementary filter as well as the Kalman filters, the precision of the position estimation at the cm level will be ensured. Thus, a system that supports each other can be created for indoor navigation. Tests of the visual inspection system will primarily be carried out in the Gazebo simulation environment. Thanks to the Gazebo simulation system, different indoor navigation algorithms (Slam, optical flow, lidar, mapping with image processing, etc.) can be simulated and compared. In addition, in case of any malfunction on the system, the system will be able to detect this and ensure that the necessary axioms are taken. After the successful completion of the simulation tests, it is aimed to implement the system simply.