1.1 1393 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Gustav Henriks author Joachim Ferreau supervisor Pontus Giselsson supervisor Anders Robertsson supervisor Department of Automatic Control v1000253 department With a steady development of technology, the use of Model Predictive Control (MPC) has become more and more popular since the computation time has gone down. With this increase, a need for determining which MPC algorithm is good for solving a certain type of MPC problem has occurred which would facilitate the choice of algorithm and also could increase the performance. One of the determining aspects is if a solver has the possibility to terminate early (stop the algorithm before it has performed all of its iterations), for example if it is placed on an embedded system with strict real-time bounds. With the use of an MPC Benchmarking suite available at ABB Corporate Research Switzerland, the early termination of MPC algorithms has been investigated. With the usage of 19 benchmarks and 3 different solvers that uses Interior point method, Gradient descent and Active-set method a large number of results have been looked through with the help of different Machine Learning methods. The result has been classified as good or bad performance when terminated early and different models have been fitted to predict this data. From this a group of key-features have been attempted to get extracted to see if there is a possibility on beforehand to tell if a control problem and a certain solver can be early terminated. Important features that were found were mostly concerning whether the control input u was under constraint or not. Good results were especially achieved for a machine learning model based on the Active-set solver qpOASES which could give good indications on whether a certain problem could get early terminated or not. http://lup.lub.lu.se/student-papers/record/8896087/file/8896088.pdf application/pdf 536337 no 2016 eng Technology and Engineering 0280-5316 TFRT-6023 http://lup.lub.lu.se/student-papers/record/8896087/file/8896090.pdf application/pdf no 8896087 2016-12-15T08:46:38+01:00 2016-12-16T13:58:41+01:00 2016-12-16T13:58:41+01:00 1 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Kirsten Carstaedt author Daniel Görges supervisor Felix Berkel supervisor Meike Stemmann supervisor professor Anders Rantzer supervisor Department of Automatic Control v1000253 department This thesis and technical report concentrates on distributed control using a distributed model predictive scheme. The model of a two room house and three room houses is build and a distributed model predictive control (MPC) algorithm is implemented in order to reach specified room temperatures with minimized energy effort in each room. For reference tracking Target Calculation and the delta input scheme are used. The MPC optimization problem is solved at each time step through an iterative method, where the number of iterations is reduced through a stopping criterion guaranteeing stability and a prespecified amount of performance and feasibility. The optimization problem is divided up into subproblems, where each subproblem takes less computational effort than the central optimization problem. Due to the possibility of coupling between subsystems, communication between the subsystems is needed. The reference values are reached and iterations needed to solve the optimization are reduced with the stopping condition. This method saves computing time and gives privacy to each subsystem, since only required information is communicated. Also the subsystems get less susceptible to the failure of one coupled subsystem, since if one subsystem fails, the others could go on. But, due to the needed communication, t h i s method is more suitable for large systems with sparse coupling. For a small system, or too much coupling the communication effort will get to high. http://lup.lub.lu.se/student-papers/record/8775922/file/8775934.pdf application/pdf 722160 no 2016 eng Technology and Engineering 0280-5316 TFRT-5998 http://lup.lub.lu.se/student-papers/record/8775922/file/8775963.pdf application/pdf no 8775922 2016-02-26T11:10:34+01:00 2016-02-28T16:54:14+01:00 2016-02-28T16:54:14+01:00 2 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 David Bergström author Robert Göransson author Simon Yngve supervisor PhD Anton Cervin supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department Model-Based Design is a development method that is becoming popular to use when creating control systems. In this thesis a demonstration of the advantages of using this method is made for Combine Control Systems AB. The 3D simulation software IndustrialPhysics is used to represent a real process in form of a gantry crane. A controller for this crane is developed in Simulink and Model-in-the-Loop (MiL) testing is done together with the 3D model. C code is then generated from the controller and transferred to a PLC. A control panel with buttons is connected to the PLC and Hardware-in-the-Loop (HiL) testing is done together with the 3D model. The result of the thesis is a working HiL rig ready to be used on technical fairs to demonstrate the capabilities of the Model-Based Design method. http://lup.lub.lu.se/student-papers/record/8776530/file/8776533.pdf application/pdf 5206448 no 2016 eng Technology and Engineering 0280-5316 TFRT-5999 http://lup.lub.lu.se/student-papers/record/8776530/file/8776535.pdf application/pdf no 8776530 2016-02-26T13:48:50+01:00 2016-02-28T16:51:31+01:00 2016-02-28T16:51:31+01:00 3 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Matthieu Lucke author Margret Bauer supervisor professor Charlotta Johnsson supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department The integration of multiple data sources and the convergence of process control systems and business intelligence layer such as the enterprise resource planning (ERP) are paving the way for important progress in plant operation optimization.<br /> <br /> Numerous companies offer “Analytics Services” to leverage this newly available mine of data but applications still appear to be limited to certain specific types of large plants. Key Performance Indicators (KPIs) is arguably the most used approach to make sense of large and complex systems in a wide variety of fields and its applicability to industrial operations is more and more common, to the extent that standardization of KPIs has become a major topic for the International Organization for Standardization (ISO). While the KPI standard ISO 22400 focuses on KPIs for manufacturing operations management at the plant level, the scope of this thesis is to bring it to the first layer of the control system: the equipment.<br /> <br /> In addition to being part of the quest for operational excellence and energy efficiency, bringing KPIs to the asset level is an important step towards integration of the different layers of the automation pyramid, integrating in particular control and scheduling. Developed within the frame of the new generation of Operations Management Software for the process industries, this work presents a case study on the most widely used assets in the field – pumps, based on operational data of different plants in the Oil &amp; Gas and Chemicals industries. http://lup.lub.lu.se/student-papers/record/8776621/file/8776629.pdf application/pdf 4994028 no 2016 eng Technology and Engineering 0280-5316 TFRT-6001 http://lup.lub.lu.se/student-papers/record/8776621/file/8776631.pdf application/pdf no 8776621 2016-02-26T14:24:49+01:00 2016-02-28T16:50:26+01:00 2016-02-28T16:50:26+01:00 4 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Sara Gunnarsson author Maria Henningsson supervisor Anders Robertsson supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department The Modelica language is an object-oriented, equation-based language used for modeling and design in many different domains and for various physical applications. In order to exchange models both between Modelicabased tools and other tools the Functional Mock-up Interface (FMI) has emerged. This interface enables exchange of virtual models between manufacturers and/or divisions within a company. The industry is moving towards increased development of virtual models, both for understanding of systems but also because it lowers the development costs.<br /> <br /> In this thesis a proof-of-concept of the newly released FMI support in Automation Studio is tested. With a physical model of a reaction wheel pendulum as base, a virtual model is designed in the Modelica-based simulation tool Dymola. A reaction wheel pendulum is a pendulum equipped with a motorized wheel, which is run in different directions in order to swing up the pendulum and to be able to balance it at its highest point. The Dymola model is converted into a Functional Mock-up Unit (FMU), which is the implementation of the FMI. Since not all values of the pendulum are known, an estimation script is written and run in JModelica.org. The FMU is then modified with these estimated values and tested with the controller included together with the reaction wheel pendulum in Simulink, in order to verify that these work together. Lastly the FMU is imported into Automation Studio together with the controller, imported with Automation Studio Target for Simulink, and simulated together there.<br /> <br /> The workflow of the thesis was successful and the controller managed to control the FMU in Automation Studio. http://lup.lub.lu.se/student-papers/record/8776878/file/8776886.pdf application/pdf 901382 no 2016 eng Technology and Engineering 0280-5316 TFRT-6002 http://lup.lub.lu.se/student-papers/record/8776878/file/8776888.pdf application/pdf no 8776878 2016-02-26T16:00:54+01:00 2016-02-28T16:48:46+01:00 2016-02-28T16:48:46+01:00 5 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Sebastian Brand author Niklas Nilsson author Martin Holmstrand supervisor Anders Robertsson supervisor Rolf Johansson supervisor Department of Automatic Control v1000253 department Today’s industrial robots are highly repeatable, but need to be calibrated to improve their absolute accuracy. This calibration can be done on many of the robot properties such as kinematic parameters, joint friction or bending stiffness. This thesis explores a calibration procedure for the kinematic parameters, that use a specialized piece of<br /> hardware - the double ball-bar.<br /> The double ball-bar restricts the motion of the robot to a spherical surface. Sensors were added to the ball-bar joints, which made it possible to use the forward kinematic homogeneous transformation matrix from both the robot side and the ball-bar side, the matrices can be compared to each other and the parameters of the robot can be identified using a non-linear least-squares minimization algorithm.<br /> The calibration proved promising in simulations and showed an increased robustness to error sources such as white noise and fluctuations of the gear ratio found in cycloid drives. It also provided an improved identification of the robot parameters compared to the calibration done using the sensor-less double ball-bar.<br /> In experiments the identification showed some improvement in the identification over using the sensor-less double ball-bar, but also that the method needs to be further improved to be able to produce satisfactory calibration results. http://lup.lub.lu.se/student-papers/record/8898906/file/8898907.pdf application/pdf 16608700 no 2016 eng Technology and Engineering 0280-5316 TFRT-6017 http://lup.lub.lu.se/student-papers/record/8898906/file/8898909.pdf application/pdf no 8898906 2017-01-13T11:05:29+01:00 2017-01-16T13:39:10+01:00 2017-01-16T13:39:10+01:00 6 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Rebecka Erntell author Mats Iderup supervisor Björn Strandmark supervisor Noroz Akhlagi supervisor Martina Maggio supervisor Anders Robertsson supervisor Department of Automatic Control v1000253 department The aim of this thesis was to explore and evaluate current methods for building tiny, low-power audio recording systems for wireless applications. Short audio clips should be recorded, transferred to a smartphone and played back. Optimization of the different parts should be explored and a prototype for usage in battery powered devices with limited space should be built.<br /> System requirements were defined in terms of audio quality, physical size and energy consumption. Additionally, the price should be kept low and the system should be reliable and stable, without notable delays.<br /> The system was divided into six focus areas; microphones, preamplifiers, signal processing, transmission, software design and energy consumption feedback. Each of those was explored theoretically and by practical prototyping, testing and tuning.<br /> Low energy consumption mainly proved to compromise audio quality, system speed and development time. Signal processing and transmission, especially range, speed and protocol overhead, played major roles.<br /> Size constraints did not affect performance very much, except for reduced battery capacity and microphone positioning options. Decent sound quality could be achieved with small size and low cost and signal processing could do much to keep energy consumption down and the perceived sound quality up. http://lup.lub.lu.se/student-papers/record/8898949/file/8898952.pdf application/pdf 4180408 no 2016 eng Technology and Engineering 0280-5316 TFRT-6024 http://lup.lub.lu.se/student-papers/record/8898949/file/8898954.pdf application/pdf no 8898949 2017-01-13T14:08:33+01:00 2017-01-16T13:46:42+01:00 2017-01-16T13:46:42+01:00 7 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Farid Alijani author Martin Karlsson supervisor Björn Olofsson supervisor Anders Robertsson supervisor Department of Automatic Control v1000253 department Docking of mobile robots requires precise position measurements in relation to the docking platform to accomplish the task successfully. Besides, the pose estimation of the robot with the sensors in an indoor environment should be accurate enough for localization and navigation toward the docking platform. However, the sensor measurements are entitled to disturbances and measurement uncertainties. In this thesis, sensor integration is exploited to decrease the measurement errors and increase the accuracy of the docking.<br /> The first approach in the autonomous docking of a mobile robot considered in this thesis is to use the already built-in laser scanner sensor to examine the feasibility of the precise docking with several experiments employing different markers.<br /> The second approach is a vision-based control method with a marker mounted on the docking platform, identified as the target. The vision-feedback control system evaluates the current position of the robot relative to the target in real time to compute the velocity commands for the actuators to reach the docking platform. The versatility of the markers is investigated in this method.<br /> The final approach is a Reinforcement Learning (RL) framework to investigate and compare the optimality of the docking with the vision-based control method. In the RL approach training is handled in a simulation environment with the reward distribution.<br /> The obtained results of the approaches are evaluated in experiments based on the desired docking behavior with respect to the trajectory and time. The control design with a real-time vision system demonstrates the capabilities of this approach to conduct accurate docking of the mobile robot starting in different configurations. http://lup.lub.lu.se/student-papers/record/8898960/file/8898961.pdf application/pdf 1783412 no 2017 eng Technology and Engineering 0280-5316 TFRT-6018 http://lup.lub.lu.se/student-papers/record/8898960/file/8898963.pdf application/pdf no 8898960 2017-01-13T14:42:57+01:00 2017-01-16T13:48:20+01:00 2017-01-16T13:48:20+01:00 8 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Marcus Greiff author Fredrik Bagge Carlson supervisor Anders Robertsson supervisor Professor Bo Bernhardsson supervisor Department of Automatic Control v1000253 department The master thesis seeks to develop a control system for the Crazyflie 2.0 unmanned aerial vehicle to enable aggressive and autonomous flight. For this purpose, different rigid-body models are considered, differing primarily in their parametrisation of rotation. The property of differential flatness is explored and several means of parametrising trajectories in at output space are implemented. A new method of rotor control with parameter estimation is developed and geometric controllers are implemented for rigid-body control. Finally, state estimation is accomplished through a scalar-update extended Kalman filter, where information from the internal measurement unit is fused with positional information from camera systems, ultra-wide band systems, optical flow measurements and laser ranging measurements. Capable of sustaining flight with any combination of the previously mentioned sensors, the real-time implementation is showcased using polynomial motion-planning to avoid known obstacles. http://lup.lub.lu.se/student-papers/record/8905295/file/8905299.pdf application/pdf 6635965 no 2017 eng Technology and Engineering 0280-5316 TFRT-6026 http://lup.lub.lu.se/student-papers/record/8905295/file/8905301.pdf application/pdf no 8905295 2017-03-31T13:15:26+02:00 2017-05-24T11:56:43+02:00 2017-05-24T11:56:43+02:00 9 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Henrik Normann author Anna Åstrand supervisor Pontus Nordin supervisor professor Charlotta Johnsson supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department This thesis focuses on the growing demand for traceability in continuous processes within the food and beverage<br /> industries. Three market-leading companies have contributed valuable input to this thesis to identify common interests. Using modern technology, a demo application has been developed to show the power in gathering several data sources and traceability into one graphical user interface, GUI.<br /> The application stands on a strong theoretical ground by taking several international standards, needs assessment and user interface theory into account. By doing so the final application emits a strong understanding of the current traceability needs and trends. This thesis was carried out together with ABB in Malmö. http://lup.lub.lu.se/student-papers/record/8905706/file/8905708.pdf application/pdf 822690 no 2017 eng Technology and Engineering 0280-5316 TFRT-6028 http://lup.lub.lu.se/student-papers/record/8905706/file/8905710.pdf application/pdf no 8905706 2017-04-12T11:14:52+02:00 2017-05-24T11:55:45+02:00 2017-05-24T11:55:45+02:00 10 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Johan Kellerth Fredlund author Kenan Sadik Sulejmanovic author Michael Bano supervisor Professor Bo Bernhardsson supervisor Anders Robertsson supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department This thesis explored the path following applications of autonomous driving, where the purpose is to sense the environment and navigate to a goal without human intervention. Autonomous driving enables safer journeys by removing human error in driving, as well as reducing fuel consumption and driving time by optimizing the engine and brake actuation. In 2017 it is a well researched topic in the academic and industrial world.<br /> The scope of this work was to create a reliable and efficient path following solution which enabled a robot car to accurately traverse along any given path. The objective was therefore to minimize path error with the additional objective to minimize the traversal time.<br /> The method used to realize the objective statement was Model Predictive Contouring Control (MPCC). MPCC uses optimization based prediction to smoothly control the car along a reference path, where actuator constraints and the objective are both handled in the optimization problem. The reference was also defined as a geometric function instead of a specified trajectory of coordinates. With the addition of objective weights it was possible to choose the relative importance of the two objectives, path accuracy and<br /> minimizing time. The resulting problem formulation, which was heavily nonlinear, was linearized to reduce computation time and solved using FORCES Pro. Simulations where made in Simulink, while real-time execution of the MPCC were achieved with a robot prototype, Robotics Operating System (ROS) and Matlab.<br /> The result suggested that MPCC was able to operate efficiently in a real time environment. Manipulating the objective weights resulted in predictable operation, which makes it possible for a user to control whether faster traversal or tracking performance is desired. It can be concluded that MPCC is a viable control method for autonomous driving, in particular for autonomous racing purposes. Additionally, the simulation results regarding the non-linear MPCC showed a surprisingly fast solve time, which suggest that it would be feasible for on-line driving using the solver FORCES Pro. http://lup.lub.lu.se/student-papers/record/8906188/file/8906189.pdf application/pdf 2675567 no 2017 eng Technology and Engineering 0280-5316 TFRT-6029 http://lup.lub.lu.se/student-papers/record/8906188/file/8906191.pdf application/pdf no 8906188 2017-04-26T11:23:37+02:00 2017-05-24T11:53:56+02:00 2017-05-24T11:53:56+02:00 11 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Joel Assarsson author Simon Thoulouis author Bo Wennberg supervisor Professor Tore Hägglund supervisor Rolf Johansson supervisor Department of Automatic Control v1000253 department The use of computer models in research and development is a powerful tool to make the engineering process more effective. By providing accurate simulations, the need for working prototypes and extensive testing is reduced. This Master’s Thesis investigates the Gambro WRO 300 H reverse osmosis water purification system with the aim of creating a working and easily adaptable computer model of the<br /> machine. For this purpose, the physical network modeling approach provided by the MathWorks Simscape environment has been used, and the experimental data from the machine is fitted to an a priori model of the reverse osmosis process - the solution-diffusion model. The result is a Simscape model of the WRO 300 H as well as four experimental model designs implemented using the same custom Simscape library. While the models do capture the overall behaviour of the WRO 300 H and provide interesting insights in the experimental designs at a qualitative level, there are significant improvements that could be made to enhance the accuracy of the membrane model by modeling its internal physical phenomena. http://lup.lub.lu.se/student-papers/record/8906434/file/8906435.pdf application/pdf 6409409 no 2017 eng Technology and Engineering 0280-5316 TFRT-6027 http://lup.lub.lu.se/student-papers/record/8906434/file/8906437.pdf application/pdf no 8906434 2017-05-04T13:22:32+02:00 2017-05-24T11:52:36+02:00 2017-05-24T11:52:36+02:00 12 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Lloyd Tran author David Barbulovic author Anders Robertsson supervisor Rolf Johansson supervisor Department of Automatic Control v1000253 department This thesis investigated how one can build a PowerLink network and Control Nodes that act as, e.g., sensor, actuators, etc., together with B&amp;R:s products. Establishing different PowerLink networks was also investigated such as PLC connected to Raspberry Pi2. Configuration and step-by-step introduction of what PowerLink is and how it works is also mentioned in this thesis. In addition to creating PowerLink networks and nodes, this thesis also looked into implementing OpenCV for image processing with a camera on a Raspberry Pi2 connected to a B&amp;R PLC.<br /> <br /> PowerLink is a deterministic Ethernet-based protocol. It works by having a Managing Node that acts as the moderator for the network. Then there are the Control Nodes, which are slaves nodes in the network. The result showed that it is possible to create control nodes that can be controlled and read through PowerLink on a PC or PLC. The PowerLink camera node that was created was implemented on Raspberry Pi2 with OpenCV. The idea behind<br /> this was to show that it is possible to easily setup a PowerLink node. One of the problems that was experienced was that the PowerLink stack and OpenCV code that was to be implemented was written in C/C++, but this was solved by establishing a Server/Client communication between the two programs. The Raspberry Pi2 limited hardware, which made it not suitable for high performance systems.<br /> <br /> This thesis showed that it is possible to create sensors, control and directly test them from a PC or implement them on a industrial PLC. For the users it is greatly beneficial due to the fact that a control node can execute code, e.g., filter measurements, before sending data to the PLC. http://lup.lub.lu.se/student-papers/record/8937349/file/8937352.pdf application/pdf 3761497 no 2018 eng Technology and Engineering 0280-5316 TFRT-6048 http://lup.lub.lu.se/student-papers/record/8937349/file/8937354.pdf application/pdf no 8937349 2018-03-14T15:02:42+01:00 2018-06-29T12:25:30+02:00 2018-06-29T12:25:30+02:00 13 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Jonatan Ekelund author Jasper Gundersen supervisor Anders Robertsson supervisor Martina Maggio supervisor Department of Automatic Control v1000253 department This master thesis project has developed a method for a six-legged robot to traverse uneven terrains. From a previous project, the robot already possessed basic locomotion which had to be extensively redesigned. Several new subsystems have been introduced, such as a continuous balancing system, a system estimating the slope of the ground the robot is walking on, and method for detecting when the robot’s legs hits the ground.<br /> This project designed a system that lets the robot balance its body as it moved using an inertial measurement unit. Several different controllers have been tested and compared. A lot of work has gone into gathering data and identifying how different controllers and sampling rates help the system. <br /> To move across uneven terrain it is important to estimate the tilt of the ground that the robot is moving on. This estimate has been made by integrating the angles from the inertial measurement unit. However, a superior solution using distance measuring sensors has been developed in a simulated environment. The distance measuring sensors also made it possible to control the height of the robot more accurately than before.<br /> A large part of the project has been spent on testing and comparing different methods of detecting when the legs of the robot hit the ground. No satisfactory solution was found using buttons or force sensitive resistors. Instead the legs’ movement deviations were used to detect when a leg touched the ground. This was then successfully implemented in simulation.<br /> Several new features, unrelated to the balancing but improving the movements of the hexapod, have been developed and implemented.<br /> Also, a mathematical model for the maximum speed of hexapod walking on sloping ground has been designed. http://lup.lub.lu.se/student-papers/record/8932396/file/8932398.pdf application/pdf 13736981 no 2018 eng Technology and Engineering 0280-5316 TFRT-6047 http://lup.lub.lu.se/student-papers/record/8932396/file/8932400.pdf application/pdf no 8932396 2018-01-16T13:14:12+01:00 2018-02-27T13:04:15+01:00 2018-02-27T13:04:15+01:00 14 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 John Kroon author Felix Matthies supervisor Prof Karl-Erik Årzén supervisor PhD Anton Cervin supervisor Department of Automatic Control v1000253 department Driving comfort is subjective; every driver has their set of preferences on how a car should behave. As the ability to control and suppress unwanted disturbances has increased with increased technological abilities, the demands on driving comfort in modern cars also rise. To account for these new demands, the testing and evaluation of the driving comfort becomes increasingly important for car manufacturers.<br /> <br /> This thesis focuses on how to implement different benchmark tests of an automatic gearbox into the Drivebox, a tool developed at IAV that can act as the accelerator pedal in a car, and the creation of an interface between the Drivebox and the test database Caltet. The final goal with the new implementation of the Drivebox is an easier and more repeatable execution of benchmark tests. In order to verify that the interface and the implemented tests work, the test were first simulated with the help of a computer and then performed in a vehicle. The results from both the simulation and the tests in the vehicle show that it is possible to control and execute the benchmark tests using this new interface with the Drivebox. http://lup.lub.lu.se/student-papers/record/7989464/file/7989473.pdf application/pdf 1978389 no 2015 eng Technology and Engineering 0280-5316 ISRN LUTFD2/TFRT--5989--SE http://lup.lub.lu.se/student-papers/record/7989464/file/7989475.pdf application/pdf no 7989464 2015-09-23T14:19:15+02:00 2015-09-25T09:45:44+02:00 2015-09-25T09:45:44+02:00 15 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Anton Norell author Oscar Linde author Marcus Johansson supervisor Rickard Andersson supervisor Martina Maggio supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department A software system composed of applications running on embedded devices could be hard to monitor and debug due to the limited possibilities to extract information about the complex process interactions. Logging and monitoring the systems behavior help in getting an insight of the system status. The information gathered can be used for improving the system and helping developers to understand what caused a malfunctioning behavior. This thesis explores the possibility of implementing an Event Sniffer that runs on an embedded Linux device and monitors processes and overall system performance to enable mapping between system usage and load on certain parts of the system. It also examines the use of data mining to process the large amount of data logged by the Event Sniffer and with this find frequent sequential patterns that cause a bug to affect the system’s performance. The final prototype of the Event Sniffer logs process cpu usage, memory usage, process function calls, interprocess communication, system overall performance and other application specific data. To evaluate the data mining of the logged information a bug pattern was planted in the interprocess communication, that caused a false malfunctioning. The data mining analysis of the logged interprocess communication was able to find the planted bug-patterna that caused the false malfunctioning. A search for a memory leak with the help of data mining was also tested by mining function calls from a process. This test found sequential patterns that was unique when the memory increased. http://lup.lub.lu.se/student-papers/record/7860324/file/7860337.pdf application/pdf 599553 no 2015 eng Generalized Sequential Pattern Debugging Embedded Linux Logging Data mining Technology and Engineering 0280-5316 ISRN LUTFD2/TFRT--5988--SE http://lup.lub.lu.se/student-papers/record/7860324/file/7860339.pdf application/pdf no 7860324 2015-09-07T15:56:02+02:00 2015-09-11T09:39:08+02:00 2015-09-11T09:39:08+02:00 16 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Stefan Olsson author Erik Lindmark supervisor PhD Anton Cervin supervisor Prof Karl-Erik Årzén supervisor Department of Automatic Control v1000253 department Alten AB has a technology demonstator in the form of a motorized and camera equipped large scale labyrinth game. The ball position is controlled by a ABB industrial PLC connected with Android tablets for user interface and a camera as a sensor for the ball position. This thesis demonstrates the ability to place a wide angle camera inside the cabinet, correcting the lens distortion caused by the wide angle lens and detect the ball with the use of a circular Hough transform. A path is also generated from the ball position to any position of the maze by capturing an image from the camera, generating a map for subsequent pathfinding, using an improvement of the Dijkstra’s pathfinding algorithm named Theta*. It further demonstrates the feasibility of using the computing power of the camera for both pathfinding and ball positioning. http://lup.lub.lu.se/student-papers/record/7866384/file/7866413.pdf application/pdf 3035223 no 2015 eng Technology and Engineering 0280-5316 ISRN LUTFD2/TFRT--5964--SE http://lup.lub.lu.se/student-papers/record/7866384/file/7866422.pdf application/pdf no 7866384 2015-09-14T13:39:22+02:00 2015-09-16T13:12:17+02:00 2015-09-16T13:12:17+02:00 17 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Patrik Önnheim author Hampus Isaksson author Gunnar Lindstedt supervisor Anders Robertsson supervisor Rolf Johansson supervisor Department of Automatic Control v1000253 department The purpose of this report is to evaluate a direct driven permanent magnet motor in high accuracy position and low speed operation. Actuation in this case is usually accomplished by stepping motors combined with belts and pulleys. High accuracy positioning is considered to be within 0.1 degrees and low speed 0.05 degrees per second, while at the same time have a 180 degree step response within 0.5 second.<br /> <br /> A model is derived of the motor along with methods for model parameter identification. This model is the basis for simulation of the motor in closed loop control. A prototype is developed in order to prove the validity of the results made by simulations. Experiments on the prototype resulted in two control methods, namely field oriented control and synchronous control.<br /> <br /> Conclusions drawn from the projects are as follows. The simulations do mirror the inherent problems with the permanent magnet motor. The prototype developed for the project is functioning and highly capable. Field oriented control was unable to meet the specified requirements. However, combined with iterative learning control the performance was improved significantly. Synchronous control satisfied most of the requirements, although its responsiveness and low efficiency are possible areas of improvement in future research. http://lup.lub.lu.se/student-papers/record/7866840/file/7866867.pdf application/pdf 740021 no 2015 eng Technology and Engineering 0280-5316 ISRN LUTFD2/TFRT--5963--SE http://lup.lub.lu.se/student-papers/record/7866840/file/7866869.pdf application/pdf no 7866840 2015-09-14T16:15:53+02:00 2015-09-16T13:07:53+02:00 2015-09-16T13:07:53+02:00 18 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Martin Vikström Morin author PhD Anton Cervin supervisor Professor Bo Bernhardsson supervisor Department of Automatic Control v1000253 department This thesis covers the optimal control of stochastic systems with coarsely quantised measurements. A particle filter approach is used both for the estimation and control problem. Three main families of particle filters are examined for state estimation, standard SIR filters, SIR filters with generalised sampling and auxiliary filters. A couple of different proposal distributions and weight functions were examined for the generalised SIR and auxiliary filter respectively. The choice of proposal distribution had the greatest impact on performance but the unrivalled best filter was achieved with a combination of generalised sampling and the auxiliary particle filter. For the problem of control the particle filter was used for cost-to-go evaluation by forward simulation in time. Simplifications of the full dynamic programming problem were done by reducing the time horizon resulting in M-measurement feedback policies and a new M-measurement cost feedback policy. One-measurement feedback and M-measurement cost feedback was examined for M 4 and although probing behaviour was observed none of the examined controllers managed to outperform a certainty equivalent controller. http://lup.lub.lu.se/student-papers/record/8915326/file/8915331.pdf application/pdf 1471056 no 2017 eng Technology and Engineering 0280-5316 TFRT-6031 http://lup.lub.lu.se/student-papers/record/8915326/file/8915333.pdf application/pdf no 8915326 2017-06-14T10:29:01+02:00 2017-08-18T11:10:35+02:00 2017-08-18T11:10:35+02:00 19 info:srw/schema/1/mods-v3.3xml studentPublicationsH3 Markus Björk author Robin Levenhammar author Håkan Runvik supervisor Josefin Berner supervisor Professor Tore Hägglund supervisor Department of Automatic Control v1000253 department The PID controller is by far the most popular controller for process control in industry today [Åström and Murray, 2008]. The combination of its relatively simple usage and its satisfactory results within several areas has gained its popularity [Visioli, 2006]. However, the tuning of the controller is crucial. An automatic tuning method was developed by [Åström and Hägglund, 1984] in the mid-eighties and auto-tuners has since then been diligently used in the process industry.<br /> This thesis will focus on the implementation of an auto-tuner based on asymmetric relay feedback in the modeling language Modelica. Also, creating a good work flow and providing valuable information for the user. The auto-tuner provides an estimated low order model together with parameters for a PID controller, based on the AMIGO tuning rules.<br /> After an introduction the report begins with a chapter containing a theoretical background concerning the PID controller, tuning methods, asymmetric relay autotuners and Modelica/Dymola. This is followed by an explanation of the method used, including a step-by-step description.<br /> The report ends with chapters dealing with results, discussions and comments about future development. The result chapter is divided into different sections, each dealing with one of the processes used for testing. Every section presents the experimental results together with comparisons against other parameter set-ups. The last chapter contains a discussion regarding the overall results, both in terms of usage and performance.<br /> The results, in terms of responses to load disturbances and step changes, were satisfactory for all processes that were tested in the thesis, although some processes required more preparatory work than others. As long as the process is in steadystate when the experiment begins, the resulting PID parameters turns out to work very good in comparison to the ones being used in the models today. Despite the<br /> good performance there are smaller things that needs to be further developed. For instance, a feature that is supposed to generate two plots automatically needs to be improved. http://lup.lub.lu.se/student-papers/record/8915999/file/8916000.pdf application/pdf 1621253 no 2017 eng Technology and Engineering 0280-5316 TFRT-6032 http://lup.lub.lu.se/student-papers/record/8915999/file/8916005.pdf application/pdf no 8915999 2017-06-15T13:29:10+02:00 2017-08-18T11:09:30+02:00 2017-08-18T11:09:30+02:00 20 1.1issn any "1401-9191 0280-5316 0280-5316"issnany1401-9191 0280-5316 0280-5316