To see the other types of publications on this topic, follow the link: System identification and control.
Dissertations / Theses on the topic 'System identification and control'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'System identification and control.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Szabo, Andrew P. "System Identification and Model-Based Control of Quadcopter UAVs." Wright State University / OhioLINK, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=wright1553197265058507.
Full text
2
Li, Liangmin. "Continuous time nonlinear system identification." Thesis, University of Sheffield, 2000. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.341867.
Full text
3
Brunke, Shelby Scott. "Nonlinear filtering and system identification algorithms for autonomous systems /." Thesis, Connect to this title online; UW restricted, 2001. http://hdl.handle.net/1773/7095.
Full text
4
Haider, Usama. "Smart Maintenance using System Identification." Thesis, Högskolan i Gävle, Avdelningen för elektroteknik, matematik och naturvetenskap, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:hig:diva-30735.
Full textAbstract:
This project discusses the use of System Identification for Smart Maintenance. System Identification is the process of finding a mathematical model of a system using empirical data. The mathematical model can then be used to detect and predict the maintenance needs, which is considered as Smart Maintenance. Smart maintenance strategies have gained pretty much importance recently, since it contributes to economically sustainable production. This project uses the LAVA-framework, proposed in [1] for non-linear system identification, which has the capability of explaining the dynamics of the system very well, and at the same time follows the principle of parsimony. A nominal model is first identified using data from a system that operates under normal operating conditions, then the identified nominal model is used to detect when the system starts to deviate from normal behavior, and these deviations indicate the deteriorations in the system. Furthermore, a new Multiple Model Method which is developed in [2] using the similar idea from LAVA, is applied on the large data set of a system that operates on separate batches and units, which identifies individual model for each batch and unit, which is then used to detect the deficient units or batches and changes in the system behavior. Finally, the proposed methods are applied to two different real world industrial cases; a Heat exchanger and a Wood Moulder Machine. In the first, the purpose is to detect the dirt in a Heat Exchanger, and in the second, the goal is to detect when the tool in a Wood Moulder Machine needs to be changed.
5
Salam, Md Abdul, and Md Mafizul Islam. "Modelling and Control System Design to control Water temperature in Heat Pump." Thesis, Karlstads universitet, Avdelningen för fysik och elektroteknik, 2013. http://urn.kb.se/resolve?urn=urn:nbn:se:kau:diva-30680.
Full textAbstract:
The thesis has been conducted at Hetvägg AB and the aim is to develop a combined PID and Model Predictive Controller (MPC) controller for an air to water heat pump system that supplies domestic hot water (DHW) to the users. The current control system is PLC based but because of its big size and expensive maintenance it must be replaced with a robust controller for the heat pump. The main goal of this project has been to find a suitable improvement strategy. By constructing a model of the system, the control system has been evaluated. First a model of the system is derived using system identification techniques in Matlab-Simulink; since the system is nonlinear and dynamic a model of the system is needed before the controller is implemented. The data has been estimated and validated for the final selection of the model in system identification toolbox and then the controller is designed for the selected model. The combined PID and MPC controller utilizes the obtained model to predict the future behavior of the system and by changing the constraints an optimal control of the system is achieved. In this thesis work, first the PID and MPC controller are evaluated and their results are compared using transient and frequency response plots. It is seen that the MPC obtained better control action than the PID controller, after some tuning the MPC controller is capable of maintaining the outlet water temperature to the reference or set point value. Both the controllers are combined to remove the minor instabilities from the system and also to obtain a better output. From the transient response behavior it is seen that the combined MPC and PID controller delivered good output response with minimal overshoot, rise time and settling time.
6
Kristinsson, Kristinn. "Genetic algorithms in system identification and control." Thesis, University of British Columbia, 1990. http://hdl.handle.net/2429/29628.
Full textAbstract:
Current online identification techniques are recursive and involve local search techniques.
In this thesis, we show how genetic algorithms, a parallel, global search technique
emulating natural genetic operators can be used to estimate the poles and zeros of a dynamical system. We also design an adaptive controller based on the estimates. The algorithms are shown to be useful for continuous time parameter identifications and to be able to identify directly physical parameters of a system. Simulations and an experiment show the technique to be satisfactory and to provide unbiased estimates in presence of colored noise.<br>Applied Science, Faculty of<br>Electrical and Computer Engineering, Department of<br>Graduate
7
Tayamon, Soma. "Nonlinear system identification with applications to selective catalytic reduction systems." Licentiate thesis, Uppsala universitet, Avdelningen för systemteknik, 2012. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-186963.
Full textAbstract:
The stringent regulations on the emissions levels of heavy duty vehicles create a demand for new methods of reducing harmful emissions from the engine. In order to be able to follow these increasingly stricter legislations, complex aftertreatment systems are used. Achievement of optimal performance of these systems requires accurate models that can be used for control design. As a result, the interest in modelling and control of aftertreatment systems has increased. This thesis deals with the modelling of the nitrogen oxide (NOx) emissions from heavy duty vehicles using the selective catalyst as an aftertreatment system for its reduction. The process of the selective catalytic reduction (SCR) is nonlinear since the chemical reactions involved are highly depending on the operating point. The momentary operating point is defined by the driving profile of the vehicle which, for example, includes cold and hot engine starts, highway and urban driving. The purpose of this thesis is to investigate different methods for nonlinear system identification of SCR systems with control in mind. The first two papers contain the theoretical work of this thesis. The first paper deals with improvement of an existing recursive prediction error method (RPEM) where a more accurate discretisation algorithm was used to improve the accuracy of the estimated nonlinear model. The second paper deals with analysis of the convergence properties of the algorithm. For this analysis several conditions were formulated that link the global and local convergence properties of the algorithm to stability properties of an associated differential equation. Global convergence to a stationary point was shown. In the third paper, the RPEM is used for identification of the SCR system and finally the fourth paper a Hammerstein–Wiener model for identification of the SCR system is applied. In both these cases the black-box models could predict the NOx behaviour of the SCR system quite well. The nonlinear models were shown to describe the SCR system more accurately than linear models.
8
Liu, Xing. "System identification and prediction using neural networks." Thesis, Cardiff University, 1993. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.388229.
Full text
9
Lee, James X. "On fuzzy logic systems, nonlinear system identification, and adaptive control." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1998. http://www.collectionscanada.ca/obj/s4/f2/dsk3/ftp05/nq26881.pdf.
Full text
10
Lee, James X. (James Xiang) Carleton University Dissertation Engineering Mechanical and Aerospace. "On fuzzy logic systems, nonlinear system identification, and adaptive control." Ottawa, 1997.
Find full text
11
Wang, Limin. "Modeling and real-time feedback control of MEMS device." Morgantown, W. Va. : [West Virginia University Libraries], 2004. https://etd.wvu.edu/etd/controller.jsp?moduleName=documentdata&jsp%5FetdId=3711.
Full textAbstract:
Thesis (Ph. D.)--West Virginia University, 2004.<br>Title from document title page. Document formatted into pages; contains v, 132 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 128-132).
12
Sharma, Aman. "System Identification of a Micro Aerial Vehicle." Thesis, Luleå tekniska universitet, Rymdteknik, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-73070.
Full textAbstract:
The purpose of this thesis was to implement an Model Predictive Control based system identification method on a micro-aerial vehicle (DJI Matrice 100) as outlined in a study performed by ETH Zurich. Through limited test flights, data was obtained that allowed for the generation of first and second order system models. The first order models were robust, but the second order model fell short due to the fact that the data used for the model was not sufficient.
13
Myklebust, Andreas. "Closed Loop System Identification of a Torsion System." Thesis, Linköping University, Department of Electrical Engineering, 2009. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-17531.
Full textAbstract:
<p>A model is developed for the Quanser torsion system available at Control Systems Research Laboratory at Chulalongkorn University. The torsion system is a laboratory equipment that is designed for the study of position control. It consists of a DC motor that drives three inertial loads that are coupled in series with the motor, and where all components are coupled to each other through torsional springs.</p><p>Several nonlinearities are observed and the most significant one is an offset in the input signal, which is compensated for. Experiments are carried out under feedback as the system is marginally stable. Different input signals are tested and used for system identification. Linear black-box state-space models are then identified using PEM, N4SID and a subspace method made for closed-loop identification, where the last two are the most successful ones. PEM is used in a second step and successfully enhances the parameter estimates from the other algorithms.</p>
14
Markusson, Ola. "Model and System Inversion with Applications in Nonlinear System Identification and Control." Doctoral thesis, KTH, Signals, Sensors and Systems, 2001. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-3287.
Full text
15
Anderson, Jeremiah P. "State-space modeling, system identification and control of a 4th order rotational mechanical system." Thesis, Monterey, California : Naval Postgraduate School, 2009. http://edocs.nps.edu/npspubs/scholarly/theses/2009/Dec/09Dec%5FAnderson.pdf.
Full textAbstract:
Thesis (M.S. in Electrical Engineering)--Naval Postgraduate School, December 2009.<br>Thesis Advisor(s): Yun, Xiaoping. Second Reader: Julian, Alex. "December 2009." Description based on title screen as viewed on January 26, 2010. Author(s) subject terms: System identification, state-space, pole placement, full state feedback, observer. Includes bibliographical references (p. 91). Also available in print.
16
Tayamon, Soma. "Nonlinear System Identification and Control Applied to Selective Catalytic Reduction Systems." Doctoral thesis, Uppsala universitet, Avdelningen för systemteknik, 2014. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-229148.
Full textAbstract:
The stringent regulations of emission levels from heavy duty vehicles create a demand for new methods for reducing harmful emissions from diesel engines. This thesis deals with the modelling of the nitrogen oxide (NOx) emissions from heavy duty vehicles using a selective catalyst as an aftertreatment system, utilising ammonia (NH3) for its reduction. The process of the selective catalytic reduction (SCR) is nonlinear, since the result of the chemical reactions involved depends on the load operating point and the temperature. The purpose of this thesis is to investigate different methods for nonlinear system identification of SCR systems with control applications in mind. The main focus of the thesis is on finding suitable techniques for effective NOx reduction without the need of over dosage of ammonia. By using data collected from a simulator together with real measured data, new black-box identification techniques are developed. Scaling and convergence properties of the proposed algorithms are analysed theoretically. Some of the resulting models are used for controller development using e.g. feedback linearisation techniques, followed by validation in a simulator environment. The benefits of nonlinear modelling and control of the SCR system are highlighted in a comparison with control based on linear models of the system. Further, a multiple model approach is investigated for simultaneous control of NOx and tailpipe ammonia. The results indicate an improvement in terms of ammonia slip reduction in comparison with models that do not take the ammonia slip into account. Another approach to NOx reduction is achieved by controlling the SCR temperature using techniques developed for LPV systems. The results indicate a reduction of the accumulated NOx.
17
Mårtensson, Jonas. "Geometric analysis of stochastic model errors in system identification." Doctoral thesis, KTH, Reglerteknik, 2007. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-4506.
Full textAbstract:
Models of dynamical systems are important in many disciplines of science, ranging from physics and traditional mechanical and electrical engineering to life sciences, computer science and economics. Engineers, for example, use models for development, analysis and control of complex technical systems. Dynamical models can be derived from physical insights, for example some known laws of nature, (which are models themselves), or, as considered here, by fitting unknown model parameters to measurements from an experiment. The latter approach is what we call system identification. A model is always (at best) an approximation of the true system, and for a model to be useful, we need some characterization of how large the model error is. In this thesis we consider model errors originating from stochastic (random) disturbances that the system was subject to during the experiment. Stochastic model errors, known as variance-errors, are usually analyzed under the assumption of an infinite number of data. In this context the variance-error can be expressed as a (complicated) function of the spectra (and cross-spectra) of the disturbances and the excitation signals, a description of the true system, and the model structure (i.e., the parametrization of the model). The primary contribution of this thesis is an alternative geometric interpretation of this expression. This geometric approach consists in viewing the asymptotic variance as an orthogonal projection on a vector space that to a large extent is defined from the model structure. This approach is useful in several ways. Primarily, it facilitates structural analysis of how, for example, model structure and model order, and possible feedback mechanisms, affect the variance-error. Moreover, simple upper bounds on the variance-error can be obtained, which are independent of the employed model structure. The accuracy of estimated poles and zeros of linear time-invariant systems can also be analyzed using results closely related to the approach described above. One fundamental conclusion is that the accuracy of estimates of unstable poles and zeros is little affected by the model order, while the accuracy deteriorates fast with the model order for stable poles and zeros. The geometric approach has also shown potential in input design, which treats how the excitation signal (input signal) should be chosen to yield informative experiments. For example, we show cases when the input signal can be chosen so that the variance-error does not depend on the model order or the model structure. Perhaps the most important contribution of this thesis, and of the geometric approach, is the analysis method as such. Hopefully the methodology presented in this work will be useful in future research on the accuracy of identified models; in particular non-linear models and models with multiple inputs and outputs, for which there are relatively few results at present.<br>QC 20100810
18
Liu, Zhang. "Structured semidefinite programs in system identification and control." Diss., Restricted to subscribing institutions, 2009. http://proquest.umi.com/pqdweb?did=2026920931&sid=1&Fmt=2&clientId=1564&RQT=309&VName=PQD.
Full text
19
Barenthin, Märta. "On input design in system identification for control." Licentiate thesis, KTH, School of Electrical Engineering (EES), 2006. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-4000.
Full textAbstract:
<p>There are many aspects to consider when designing system identification experiments in control applications. Input design is one important issue. This thesis considers input design both for identification of linear time-invariant models and for stability validation.</p><p>Models obtained from system identification experiments are uncertain due to noise present in measurements. The input spectrum can be used to shape the model quality. A key tool in input design is to introduce a linear parametrization of the spectrum. With this parametrization a number of optimal input design problems can be formulated as convex optimization programs. An Achilles' heel in input design is that the solution depends on the system itself, and this problem can be handled by iterative procedures where the input design is based on a model of the system. Benefits of optimal input design are quantified for typical industrial applications. The result shows that the experiment time can be substantially shortened and that the input power can be reduced.</p><p>Another contribution of the thesis is a procedure where input design is connected to robust control. For a certain system structure with uncertain parameters, it is shown that the existence of a feedback controller that guarantees a given performance specification can be formulated as a convex optimization program. Furthermore, a method for input design for multivariable systems is proposed. The constraint on the model quality is transformed to a linear matrix inequality using a separation of graphs theorem. The result indicates that in order to obtain a model suitable for control design, it is important to increase the power of the input in the low-gain direction of the system relative to the power in the high-gain direction.</p><p>A critical issue when validating closed-loop stability is to obtain an accurate estimate of the maximum gain of the system. This problem boils down to finding the input signal that maximizes the gain. Procedures for gain estimation of nonlinear systems are proposed and compared. One approach uses a model of the system to design the optimal input. In other approaches, no model is required, and the system itself determines the optimal input sequence in repeated experiments.</p>
20
Barenthin, Märta. "On input design in system identification for control /." Stockholm : Automatic control, School of Electrical Engineering, Royal Institute of Technology, 2006. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-4000.
Full text
21
DeVilbiss, Stewart L. "System Identification for H(Infinity) Robust Control Design /." The Ohio State University, 1994. http://rave.ohiolink.edu/etdc/view?acc_num=osu1487859313345322.
Full text
22
Lan, Jing. "Gaussian mixture model based system identification and control." [Gainesville, Fla.] : University of Florida, 2006. http://purl.fcla.edu/fcla/etd/UFE0014640.
Full text
23
Zhou, Xiangrong, and 周向榮. "An integrated approach to identification and control system design." Thesis, The University of Hong Kong (Pokfulam, Hong Kong), 2000. http://hub.hku.hk/bib/B31241414.
Full text
24
Choi, Eric M. "The modeling and system identification of the Dynamics Identification and Control Experiment." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1997. http://www.collectionscanada.ca/obj/s4/f2/dsk2/ftp04/mq29363.pdf.
Full text
25
Lyzell, Christian. "Initialization Methods for System Identification." Licentiate thesis, Linköping University, Linköping University, Automatic Control, 2009. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-51688.
Full textAbstract:
<p>In the system identification community a popular framework for the problem of estimating a parametrized model structure given a sequence of input and output pairs is given by the prediction-error method. This method tries to find the parameters which maximize the prediction capability of the corresponding model via the minimization of some chosen cost function that depends on the prediction error. This optimization problem is often quite complex with several local minima and is commonly solved using a local search algorithm. Thus, it is important to find a good initial estimate for the local search algorithm. This is the main topic of this thesis.</p><p>The first problem considered is the regressor selection problem for estimating the order of dynamical systems. The general problem formulation is difficult to solve and the worst case complexity equals the complexity of the exhaustive search of all possible combinations of regressors. To circumvent this complexity, we propose a relaxation of the general formulation as an extension of the nonnegative garrote regularization method. The proposed method provides means to order the regressors via their time lag and a novel algorithmic approach for the \textsc{arx} and \textsc{lpv-arx} case is given.</p><p> </p><p>Thereafter, the initialization of linear time-invariant polynomial models is considered. Usually, this problem is solved via some multi-step instrumental variables method. For the estimation of state-space models, which are closely related to the polynomial models via canonical forms, the state of the art estimation method is given by the subspace identification method. It turns out that this method can be easily extended to handle the estimation of polynomial models. The modifications are minor and only involve some intermediate calculations where already available tools can be used. Furthermore, with the proposed method other a priori information about the structure can be readily handled, including a certain class of linear gray-box structures. The proposed extension is not restricted to the discrete-time case and can be used to estimate continuous-time models.</p><p> </p><p>The final topic in this thesis is the initialization of discrete-time systems containing polynomial nonlinearities. In the continuous-time case, the tools of differential algebra, especially Ritt's algorithm, have been used to prove that such a model structure is globally identifiable if and only if it can be written as a linear regression model. In particular, this implies that once Ritt's algorithm has been used to rewrite the nonlinear model structure into a linear regression model, the parameter estimation problem becomes trivial. Motivated by the above and the fact that most system identification problems involve sampled data, a version of Ritt's algorithm for the discrete-time case is provided. This algorithm is closely related to the continuous-time version and enables the handling of noise signals without differentiations.</p>
26
Guidi, Hernan. "Open and closed-loop model identification and validation." Pretoria : [s.n.], 2009. http://upetd.up.ac.za/thesis/available/etd-07032009-170311/.
Full text
27
Chowdhary, Mahesh. "On-line system identification for control system applications in particle accelerators." W&M ScholarWorks, 1997. https://scholarworks.wm.edu/etd/1539623898.
Full textAbstract:
Particle accelerators require a number of feedback systems in order to stabilize a variety of parameters. The Continuous Electron Beam Accelerator at Thomas Jefferson National Accelerator Facility presents a unique set of control and identification problems. This accelerator produces a continuous electron beam with energies between 0.5 and 4.0 GeV to be delivered to the experimental halls. In order to meet stringent beam quality requirements specified by the experimental halls, the position and the energy of the electron beam needs to stabilized at various locations in the accelerator.;A number of noise measurement tests were conducted at various locations in the accelerator to obtain accurate information about the amplitude and the frequency of disturbances on the beam orbit and energy. Results of these measurements indicate that the line power harmonics were the primary source of disturbance on the beam orbit and energy.;A prototype fast feedback system was implemented in the injector and the East Arc regions of the accelerator to stabilize the beam position and energy at these locations. The scheme of implementation of these systems and measurements of their performance are presented here.;These feedback systems have to operate under conditions of varying noise characteristics and changing dynamics of the systems. For the feedback systems to always perform optimally, the knowledge of time varying noise characteristics and changing system dynamics needs to be incorporated into the feedback strategy. The approach presented in this work is to perform on-line system identification using a formulation of Fast Transversal Filter (FTF) in order to extract the time varying information from input/output data of the feedback system.;A simulation test stand was developed using an analog computer to represent a continuous time system whose noise characteristics and dynamics could be changed in a controlled manner. An on-line system identification algorithm was implemented on a microprocessor similar to the ones used in the accelerator control system. Experience with the hardware-in-loop simulation for various cases of changing system dynamics and noise characteristics and the performance results of the on-line system identification algorithm operating under these conditions are presented in this dissertation.
28
Brus, Linda. "Nonlinear Identification and Control with Solar Energy Applications." Doctoral thesis, Uppsala : Acta Universitatis Upsaliensis, 2008. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-8594.
Full text
29
Khader, Shahbaz Abdul. "System Identification of Active Magnetic Bearing for Commissioning." Thesis, Uppsala universitet, Institutionen för informationsteknologi, 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-243630.
Full textAbstract:
Active magnetic bearing (AMB) is an ideal bearing solution for high performance and energy efficient applications. Proper operation of AMB can be achieved only with advanced feedback control techniques. An identified system model is required for synthesizing high performance model based controllers. System identification is the preferred method for obtaining an accurate model. Therefore, it becomes a prerequisite for the commissioning of AMB. System identification for commissioning poses some challenges and special requirements. In this thesis, system identification of AMB is approached within the context of commissioning. A procedure for identification is developed and applied to experimental data from a prototype AMB system. The identification procedure is based on the so called prediction error method, and it has been performed in the frequency domain. A linear state-space model, along with the required parameters, is successfully identified.
30
Hariri, Bassam. "Modelling and identification of S.I. engines for control system design." Thesis, University of Liverpool, 1996. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.266264.
Full text
31
Eklund, John M. "Aircraft flight control simulation using parallel cascade system identification." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1998. http://www.collectionscanada.ca/obj/s4/f2/dsk2/tape17/PQDD_0004/MQ28194.pdf.
Full text
32
Spindler, Henry C. (Henry Carlton) 1970. "System identification and optimal control for mixed-mode cooling." Thesis, Massachusetts Institute of Technology, 2004. http://hdl.handle.net/1721.1/30334.
Full textAbstract:
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 2004.<br>"September 2004."<br>Includes bibliographical references (p. 285-294).<br>The majority of commercial buildings today are designed to be mechanically cooled. To make the task of air conditioning buildings simpler, and in some cases more energy efficient, windows are sealed shut, eliminating occupants' direct access to fresh air. Implementation of an alternative cooling strategy-mixed-mode cooling-is demonstrated in this thesis to yield substantial savings in cooling energy consumption in many U.S. locations. A mixed-mode cooling strategy is one that relies on several different means of delivering cooling to the occupied space. These different means, or modes, of cooling could include: different forms of natural ventilation through operable windows, ventilation assisted by low-power fans, and mechanical air conditioning. Three significant contributions are presented in this thesis. A flexible system identification framework was developed that is well-suited to accommodate the unique features of mixed-mode buildings. Further, the effectiveness of this framework was demonstrated on an actual multi- zone, mixed-mode building, with model prediction accuracy shown to exceed that published for other naturally ventilated or mixed-mode buildings, none of which exhibited the complexity of this building. Finally, an efficient algorithm was constructed to optimize control strategies over extended planning horizons using a model-based approach. The algorithm minimizes energy consumption subject to the constraint that indoor temperatures satisfy comfort requirements. The system identification framework was applied to another mixed-mode building, where it was found that the aspects integral to the modeling framework led to prediction improvements relative to a simple model.<br>(cont.) Lack of data regarding building apertures precluded the use of the model for control purposes. An additional contribution was the development of a procedure for extracting building time constants from experimental data in such a way that they are constrained to be physically meaningful.<br>by Henry C. Spindler.<br>Ph.D.
33
Lopes, Rafael Anderson Martins. "Aircraft identification applied to closed loop control system design." Instituto Tecnológico de Aeronáutica, 2006. http://www.bd.bibl.ita.br/tde_busca/arquivo.php?codArquivo=904.
Full textAbstract:
In the present work, the influence of maneuver design for stability derivatives of aerodynamic forces and moments identification in the performance of closed loop control systems is evaluated, where two maneuvers largely used in aeronautical industry - doublet and 3211 - are compared to a maneuver of restricted use in aeronautical industry, but already well known in identification - the PRBS. This evaluation is divided in two steps: in the first, it is performed the identification with each of the maneuvers and then a crosschecking of the obtained models, computing the mean quadratic prediction error; in the second, the identified models are analytically linearized and use in a bank angle tracker. The calculated gains are then applied to the nonlinear model and the performance indices computed, as peak time and maximum overshoot, of step response with each gain set. The crosschecking showed the superior capacity of PRBS maneuver to capture the dynamics of the reference model, where the prediction error was relatively small for all models when compared to the response to doublet. With the 3211 data set, the error of the model identified with doublet was significantly larger than 3211 and PRBS models error, that had similar performance, however lightly better for the PRBS. With the PRBS data set, the PRBS error was relatively small, while the doublet and 3211 models error was relatively large. This result shows the direct relation of the maneuvers frequency spectra and the excitation persistence characteristic in the identified models.With an exigent performance requirement to the bank angle tracker, it was possible to expose the differences between the models and verify the influence of these differences in the closed loop response. In an equivalent form, as observed in the crosschecking analysis, the control system designed with the PRBS model showed performance indices closer to the target in comparison to the control system designed with the 3211 model, that was significatively better than the system designed with the doublet model. The closed loop amplified the difference between the models, and depending on the control system structure, these differences can be still larger. Finally, a practical application on aircraft identification and control is presented and discussed.
34
González, Rodrigo A. "Consistency and efficiency in continuous-time system identification." Licentiate thesis, KTH, Reglerteknik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-273176.
Full textAbstract:
Continuous-time system identification deals with the problem of building continuous-time models of dynamical systems from sampled input and output data. In this field, there are two main approaches: indirect and direct. In the indirect approach, a suitable discrete-time model is first determined, and then it is transformed into continuous-time. On the other hand, the direct approach obtains a continuous-time model directly from the sampled data. In both approaches there exists a dichotomy between discrete-time data and continuous-time models, which can induce robustness issues and complications in the theoretical analysis of identification algorithms. These difficulties are addressed in this thesis. First, we consider the indirect approach to continuous-time system identification. For a zero-order hold sampling mechanism, this approach usually leads to a transfer function estimate with relative degree one, independent of the relative degree of the strictly proper true system. Inspired by the indirect prediction error method, we propose an indirect-approach estimator that enforces the desired number of poles and zeros in the continuous-time transfer function estimate, and show that the estimator is consistent and asymptotically efficient. A robustification of this method is also developed, by which the estimates are also guaranteed to deliver stable models. In the second part of the thesis, we analyze asymptotic properties of the Simplified Refined Instrumental Variable method for Continuous-time systems (SRIVC), which is one of the most popular direct identification methods. This algorithm applies an adaptive prefiltering to the sampled input and output that requires assumptions on the intersample behavior of the signals. We present a comprehensive analysis on the consistency and asymptotic efficiency of the SRIVC estimator while taking into account the intersample behavior of the input signal. Our results show that the SRIVC estimator is generically consistent when the intersample behavior of the input is known exactly and subsequently used in the implementation of the algorithm, and we give conditions under which consistency is not achieved. In terms of statistical efficiency, we compute the asymptotic Cramér-Rao lower bound for an output error model structure with Gaussian noise, and derive the asymptotic covariance of the SRIVC estimates. We conclude that the SRIVC estimator is asymptotically efficient under mild conditions, and that this property can be lost if the intersample behavior of the input is not carefully accounted for in the SRIVC procedure. Moreover, we propose and analyze the statistical properties of an extension of SRIVC that is able to deal with input signals that cannot be interpolated exactly via hold reconstructions. The proposed estimator is generically consistent for any input reconstructed using zero or first-order-hold devices, and we show that it is generically consistent for continuous-time multisine inputs as well. Comparisons with the Maximum Likelihood technique and an analysis of the iterations of the method are provided, in order to reveal the influence of the intersample behavior of the output and to propose new robustifications to the SRIVC algorithm.<br><p>QC 20200511</p>
35
Juillet, Fabien. "Control of convection-dominated flows." Palaiseau, Ecole polytechnique, 2013. http://pastel.archives-ouvertes.fr/docs/01/00/94/63/PDF/PhD_Juillet_Fabien.pdf.
Full textAbstract:
Cette thèse s'intéresse à la mise en œuvre d'une technique de contrôle d'écoulements, d'un point de vue à la fois numérique et expérimental. L'objectif de cette technique est la réduction de perturbations au sein d'écoulements dominés par la convection. Dans ce but, trois aspects sont développés au sein d'un première partie. On observe tout d'abord que dans de tels systèmes l'information voyage essentiellement vers l'aval. Pour cette raison les perturbations doivent être mesurées le plus tôt possible, en plaçant les capteurs en amont. Cette idée intuitive est étudiée quantitativement en introduisant le concept de " longueur de visibilité ". Ensuite, une description de l'écoulement est obtenue à l'aide de technique d'identification de systèmes. Ces outils présentent l'avantage de construire des modèles en se fondant uniquement sur des données accessibles au sein d'une expérience. Enfin, une approche de contrôle du type feed-forward étant particulièrement appropriée pour ces écoulements, une comparaison théorique et numérique avec la théorie classique LQG (Linear Quadratic Gaussian) est menée. Dans une seconde partie, ces trois aspects sont pris en compte dans une procédure d'identification et de contrôle qui est simplifiée de manière à faciliter une mise en place expérimentale. En particulier, les réponses impulsionnelles du système sont identifiées puis utilisées directement dans le calcul de la loi de contrôle. Cette technique repose alors uniquement sur de simples minimisations par moindres carrés et présente l'avantage d'être fondée sur des quantités aux interprétations physiques claires, telles que des vitesses de convection ou des fréquences caractéristiques. Enfin, dans une dernière partie, la procédure est appliquée expérimentalement au contrôle de perturbations dans une écoulement de Poiseuille à Re =870. Dans cette expérience, l'amplitude du signal mesuré par un capteur objectif a pu être réduite de 45%<br>In this thesis, a flow control procedure is developed numerically and is then implemented experimentally. The purpose of this procedure is to reduce the amplitude of perturbations in convection dominated flows. To design such a technique three aspects are analyzed in a first part. Since information in convection-dominated flows essentially travel downstream, incoming perturbations are better described by placing sensors upstream. This intuitive idea is studied quantitatively by introducing the concept of visibility length. In addition, a description of the flow dynamics is obtained using system identification techniques. These tools have the advantage of providing models based solely on experimentally accessible data and are therefore directly applicable to real flows. Finally, a feed-forward control approach is found to be most appropriate and a comparison with the classical linear quadratic gaussian technique is presented from numerical and theoretical point of views. In a second part, these three aspects are then taken into account in the design of a feed-forward identification and control procedure, which is then simplified to be more amenable to practical implementations in experiments. In particular, the system impulse responses are first identified, and are then directly used for the computation of the control law. Hence, the technique only relies on simple least-squares minimizations and has the advantage of manipulating quantities that have clear physical meanings, such as perturbation convective speeds and characteristic frequencies. Thus, in a last part, the control procedure is applied experimentally to the quenching of natural disturbances in a plane channel flow at Re = 870. Results show that the magnitude of the signal recorded by the objective sensor can be reduced by up to 45%
36
Kramer, Boris Martin Josef. "Model and Data Reduction for Control, Identification and Compressed Sensing." Diss., Virginia Tech, 2015. http://hdl.handle.net/10919/75179.
Full textAbstract:
This dissertation focuses on problems in design, optimization and control of complex, large-scale dynamical systems from different viewpoints. The goal is to develop new algorithms and methods, that solve real problems more efficiently, together with providing mathematical insight into the success of those methods. There are three main contributions in this dissertation.
In Chapter 3, we provide a new method to solve large-scale algebraic Riccati equations, which arise in optimal control, filtering and model reduction. We present a projection based algorithm utilizing proper orthogonal decomposition, which is demonstrated to produce highly accurate solutions at low rank. The method is parallelizable, easy to implement for practitioners, and is a first step towards a matrix free approach to solve AREs. Numerical examples for n >= 100,000 unknowns are presented.
In Chapter 4, we develop a system identification method which is motivated by tangential interpolation. This addresses the challenge of fitting linear time invariant systems to input-output responses of complex dynamics, where the number of inputs and outputs is relatively large. The method reduces the computational burden imposed by a full singular value decomposition, by carefully choosing directions on which to project the impulse response prior to assembly of the Hankel matrix. The identification and model reduction step follows from the eigensystem realization algorithm. We present three numerical examples, a mass spring damper system, a heat transfer problem, and a fluid dynamics system. We obtain error bounds and stability results for this method.
Chapter 5 deals with control and observation design for parameter dependent dynamical systems. We address this by using local parametric reduced order models, which can be used online. Data available from simulations of the system at various configurations (parameters, boundary conditions) is used to extract a sparse basis to represent the dynamics (via dynamic mode decomposition). Subsequently, a new compressed sensing based classification algorithm is developed which incorporates the extracted dynamic information into the sensing basis. We show that this augmented classification basis makes the method more robust to noise, and results in superior identification of the correct parameter. Numerical examples consist of a Navier-Stokes, as well as a Boussinesq flow application.<br>Ph. D.
37
Silva, Seth F. "Applied System Identification for a Four Wheel Reaction Wheel Platform." DigitalCommons@CalPoly, 2010. https://digitalcommons.calpoly.edu/theses/328.
Full textAbstract:
Applied System Identification for a Four Wheel Reaction Wheel Platform
By
Seth Franklyn Silva
At the California Polytechnic State University, San Luis Obispo there is a four-wheel reaction wheel pyramidal simulator platform supported by an air-bearing. This simulator has the current capability to measure the wheel speeds and angular velocity of the platform, and with these measurements, the system identification process was used to obtain the mass properties of this simulator. A handling algorithm was developed to allow wireless data acquisition and command to the spacecraft simulator from a “ground” computer allowing the simulator to be free of induced torques due to wiring. The system identification algorithm using a least squares estimation scheme was tested on this simulator and compared to theoretical analysis. The resultant principle inertia about the z-axis from the experimental analysis was 3.5 percent off the theoretical, while the other inertias had an error of up to 187 percent. The error is explained as noise attributed to noise in the measurement, averaging inconsistencies, low bandwidth, and derivation of accelerations from measured data.
38
Merkoulova, Daniel. "Optimal Input Design by Model Predictive Control for System Identification." Thesis, KTH, Reglerteknik, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-215712.
Full textAbstract:
Mathematical models are an essential part of analysis of autonomous systemsas they ease the formulation of control laws and allow experiments tobe performed in a simulation environment. For complex systems, parts of themodel may be missing which increases uncertainties and limits practical applications.Using input-output measurements makes it possible to estimate themodel, but requires the measurements to be informative. The idea in so-calledinput design is to find an input sequence for the system such that the measurementsreveal the properties and dynamics of the true system as much aspossible. This is commonly formulated as an optimization problem.This thesis focuses on formulating an optimization algorithm for inputdesign, which is implemented as an open-loop receding horizon optimizationproblem. The problem consists of classical A-, D-, E-optimality criteria andfinite combinatorial constraints on the input signal. The resulting optimizationproblem is non-convex. Two approaches are explored to solve the problem;a combinatorial and a convex relaxation. The method is finally evaluated ina simulation environment on a second-order system and first-order linearizedwater tank system.<br>Matematiska modeller är essentiella när det kommer till autonoma systemeftersom det förenklar formuleringar av styrlagar och tillåter tester i simuleringsmiljöer.För komplexa system kan delar av modellen saknas vilket ökarosäkerheten hos modellen och begränsar praktiska tillämpningar. Genom attanvända in-ut data är det möjligt att estimera modellen men det kräver attmätningarna är av relativ kvalitet. Idén är således att finna en sekvens av indataså att dynamiken hos systemet avslöjas vilket kan formuleras som ettoptimeringsproblem.Det här arbetet fokuserar på att formulera ett optimeringsproblem somska implementeras i modell prediktiv reglering. Problemet består utav klassiskA,D, E-optimalitet med begränsade diskreta bivillkor på insignalen. Metodenevalueras i simulering på ett godtyckligt andra ordningens system och på ettlineariserat första ordningens vattentankssystem.
39
Brunell, Brent Jerome 1972. "A system identification approach to active control of thermoacoustic instabilities." Thesis, Massachusetts Institute of Technology, 2000. http://hdl.handle.net/1721.1/88832.
Full text
40
Rathnasingham, Ruben. "System identification and active control of a turbulent boundary layer." Thesis, Massachusetts Institute of Technology, 1997. http://hdl.handle.net/1721.1/10468.
Full text
41
Pietersen, Willem Hermanus. "System identification for fault tolerant control of unmanned aerial vehicles." Thesis, Stellenbosch : University of Stellenbosch, 2010. http://hdl.handle.net/10019.1/4164.
Full textAbstract:
Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2010.<br>ENGLISH ABSTRACT: In this project, system identification is done on the Modular Unmanned Aerial Vehicle
(UAV). This is necessary to perform fault detection and isolation, which is part
of the Fault Tolerant Control research project at Stellenbosch University.
The equations necessary to do system identification are developed. Various methods
for system identification is discussed and the regression methods are implemented.
It is shown how to accommodate a sudden change in aircraft parameters
due to a fault. Smoothed numerical differentiation is performed in order to acquire
data necessary to implement the regression methods.
Practical issues regarding system identification are discussed and methods for
addressing these issues are introduced. These issues include data collinearity and
identification in a closed loop.
The regression methods are implemented on a simple roll model of the Modular
UAV in order to highlight the various difficulties with system identification. Different
methods for accommodating a fault are illustrated.
System identification is also done on a full nonlinear model of the Modular UAV.
All the parameters converges quickly to accurate values, with the exception of Cl R
,
CnP and Cn A
. The reason for this is discussed. The importance of these parameters
in order to do Fault Tolerant Control is also discussed.
An S-function that implements the recursive least squares algorithm for parameter
estimation is developed. This block accommodates for the methods of applying the
forgetting factor and covariance resetting. This block can be used as a stepping stone
for future work in system identification and fault detection and isolation.<br>AFRIKAANSE OPSOMMING: In hierdie projek word stelsel identifikasie gedoen op die Modulêre Onbemande Vliegtuig.
Dit is nodig om foutopsporing en isolasie te doen wat ’n deel uitmaak van fout
verdraagsame beheer.
Die vergelykings wat nodig is om stelsel identifikasie te doen is ontwikkel. Verskeie
metodes om stelsel identifikasie te doen word bespreek en die regressie metodes is
uitgevoer. Daar word gewys hoe om voorsiening te maak vir ’n skielike verandering
in die vliegtuig parameters as gevolg van ’n fout. Reëlmatige numeriese differensiasie
is gedoen om data te verkry wat nodig is vir die uitvoering van die regressie metodes.
Praktiese kwessies aangaande stelsel identifikasie word bespreek en metodes om
hierdie kwessies aan te spreek word gegee. Hierdie kwessies sluit interafhanklikheid
van data en identifikasie in ’n geslote lus in.
Die regressie metodes word toegepas op ’n eenvoudige rol model van die Modulêre
Onbemande Vliegtuig om die verskeie kwessies aangaande stelsel identifikasie uit te
wys. Verskeie metodes vir die hantering vir ’n fout word ook illustreer.
Stelsel identifikasie word ook op die volle nie-lineêre model van die Modulêre
Onbemande Vliegtuig gedoen. Al die parameters konvergeer vinnig na akkurate
waardes, met die uitsondering van Cl R
, CnP and Cn A
. Die belangrikheid van
hierdie parameters vir fout verdraagsame beheer word ook bespreek.
’n S-funksie blok vir die rekursiewe kleinste-kwadraat algoritme is ontwikkel. Hierdie
blok voorsien vir die metodes om die vergeetfaktor en kovariansie herstelling
te implementeer. Hierdie blok kan gebruik word vir toekomstige werk in stelsel
identifikasie en foutopsporing en isolasie.
42
Demarchi, Fabio Luciano. "Modeling and identification of a fly-by-wire control system." Instituto Tecnológico de Aeronáutica, 2005. http://www.bd.bibl.ita.br/tde_busca/arquivo.php?codArquivo=185.
Full textAbstract:
This work investigates the system identification and modeling techniques applied to a fly-by-wire system for pitch control of a commercial jet aircraft. The objective of the work is to build a model based on system identification techniques and generic modeling of the system, therefore using the "grey box" approach. The identification data was obtained from experimental tests performed at Embraer "Iron Bird" laboratory. An overview on flight controls systems is presented, focusing on fly-by-wire technology. To provide the theoretical bases for the experimental identification, a review on system identification techniques is presented, together with the preliminary modeling and determination of model structure. It is further presented the identification test laboratory configuration, test procedure and results analysis using the Matlab "System Identification Toolbox". The resulting transfer function obtained from system identification process is used to identify the dynamical characteristics of the system's components (hydraulic actuator, servo-valve, electronic control). The linear model identified is therefore analyzed and validated and the non-linearities identified during the analysis are included in the final complete model.
43
Choi, Ju-Yeop. "Nonlinear system identification and control using a neural network approach." Diss., Virginia Tech, 1994. http://hdl.handle.net/10919/40199.
Full text
44
Chang, Gary Carleton University Dissertation Engineering Systems and Computer. "System identification and control of a thermo-mechanical pulping refiner." Ottawa, 1995.
Find full text
45
Chamberlain, Caleb H. "System Identification, State Estimation, and Control of Unmanned Aerial Robots." BYU ScholarsArchive, 2011. https://scholarsarchive.byu.edu/etd/2605.
Full textAbstract:
This thesis describes work in a variety of topics related to aerial robotics, including system identification, state estimation, control, and path planning. The path planners described in this thesis are used to guide a fixed-wing UAV along paths that optimize the aircraft's ability to track a ground target. Existing path planners in the literature either ignore occlusions entirely, or they have limited capability to handle different types of paths. The planners described in this thesis are novel in that they specifically account for the effect of occlusions in urban environments, and they can produce a much richer set of paths than existing planners that account for occlusions. A 3D camera positioning system from Motion Analysis is also described in the context of state estimation, system identification, and control of small unmanned rotorcraft. Specifically, the camera positioning system is integrated inside a control architecture that allows a quadrotor helicopter to fly autonomously using truth data from the positioning system. This thesis describes the system architecture in addition to experiments in state estimation, control, and system identification. There are subtleties involved in using accelerometers for state estimation onboard flying rotorcraft that are often ignored even by researchers well-acquainted with the UAV field. In this thesis, accelerometer-rotorcraft behavior is described in detail. The consequences of ignoring accelerometer-rotorcraft behavior are evaluated, and an observer is presented that achieves better performance by specifically modeling actual accelerometer behavior. The observer is implemented in hardware and results are presented.
46
Lei, Yu. "Functional Regression and Adaptive Control." Diss., Virginia Tech, 2012. http://hdl.handle.net/10919/29113.
Full textAbstract:
The author proposes a novel functional regression method for parameter estimation and adaptive control in this dissertation. In the functional regression method, the regressors and a signal which contains the information of the unknown parameters are either determined from raw measurements or calculated as the functions of the measurements. The novel feature of the method is that the algorithm maps the regressors to the functionals which are represented in terms of customized test functions. The functionals are updated continuously by the evolution laws, and only an infinite number of variables are needed to compute the functionals. These functionals are organized as the entries of a matrix, and the parameter estimates are obtained using either the generalized inverse method or the transpose method. It is shown that the schemes of some conventional adaptive methods are recaptured if certain test function designs are employed. It is proved that the functional regression method guarantees asymptotic convergence of the parameter estimation error to the origin, if the system is persistently excited. More importantly, in contrast to the conventional schemes, the parameter estimation error may be expected to converge to the origin even when the system is not persistently excited. The novel adaptive method are also applied to the Model Reference Adaptive Controller (MRAC) and adaptive observer. It is shown that the functional regression method ensures asymptotic stability of the closed loop systems. Additionally, the studies indicate that the transient performance of the closed loop systems is improved compared to that of the schemes using the conventional adaptive methods. Besides, it is possible to analyze the transient responses a priori of the closed loop systems with the functional regression method. The simulations verify the theoretical analyses and exhibit the improved transient and steady state performances of the closed loop systems.<br>Ph. D.
47
Wang, Dawei. "Robust system identification and robust model predictive control with applications to chemical engineering processes." Thesis, The University of Sydney, 2001. https://hdl.handle.net/2123/27788.
Full textAbstract:
This thesis deals with problems in two closely related research areas, system identification
(SYSID) and model predictive control (MPC), which are both important in industrial.
process applications. The robustness of SYSID and MPC is the main concern of the thesis.
The focus of the first part of the thesis is to develop Robust System Identification
Strategies that yield effective and robust estimates of system model and characteristics
under the uncertainty of a priori knowledge. The incorporation of the techniques taken
from classical system identification approaches as well as robust identification design
based on Wavelets density estimation will ensure that the estimation of system is “optimal”
in all assumptions and supposed operation conditions considered of the process error. The
conclusive result has been a robust and effective estimator that is in accordance to the
industrial requirements, in regards to practical implementation of these strategies. This
method is applied to most system models and estimation algorithms (especially on—line
recursive algorithms) and takes the conventional ones as special cases. This approach has
many innovative features, one being the design of robust estimator which is insensitive to
deviation from the various assumptions about the prediction error. This characteristic
makes a robust estimation to be suitable for most error distributions. This methodology
would then be capable of solving a large number of difficult problems in relation to
modeling a system using contaminated data facing in fields such as industry, science and
economics.
The second problem addressed in the thesis is to develop Robust Model Predictive
Control Strategies that yield effective and robust controller under the mismatch between
plant and the models employed to describe the system. A new robust MPC design method
is derived from comparison and discussion of the similarity between parameter estimation
and MPC. The incorporation of techniques taken from standard model predictive control
approaches as well as the robust controller design will ensure that the controller has better
performance regarding assumptions on the nominal model and assumed plant-model
mismatch. The result will be a robust and effective controller that is in accordance to the
industrial requirements. Some properties of the proposed robust MPC design such as
closed loop stability and performance tuning are investigated. The innovative feature of the
robust MPC design makes the controller insensitive to deviation from the various
assumptions about the nominal model, and to be suitable for most plant-model
mismatches. The proposed design can take into account the compromise between
efficiency, robustness as well as computation load. It is shown that it is an expansion of '
conventional MPC.
The practical implications of the proposed robust system identification and robust MPC
design methods are shown by means of chemical engineering process examples.
48
Alukaidey, R. A. S. "Multivariable identification and adaptive control." Thesis, Brunel University, 1989. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.384517.
Full text
49
Malmström, Magnus. "Uncertainties in Neural Networks : A System Identification Approach." Licentiate thesis, Linköpings universitet, Reglerteknik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-174720.
Full textAbstract:
In science, technology, and engineering, creating models of the environment to predict future events has always been a key component. The models could be everything from how the friction of a tire depends on the wheels slip to how a pathogen is spread throughout society. As more data becomes available, the use of data-driven black-box models becomes more attractive. In many areas they have shown promising results, but for them to be used widespread in safety-critical applications such as autonomous driving some notion of uncertainty in the prediction is required. An example of such a black-box model is neural networks (NNs). This thesis aims to increase the usefulness of NNs by presenting an method where uncertainty in the prediction is obtained by linearization of the model. In system identification and sensor fusion, under the condition that the model structure is identifiable, this is a commonly used approach to get uncertainty in the prediction from a nonlinear model. If the model structure is not identifiable, such as for NNs, the ambiguities that cause this have to be taken care of in order to make the approach applicable. This is handled in the first part of the thesis where NNs are analyzed from a system identification perspective, and sources of uncertainty are discussed. Another problem with data-driven black-box models is that it is difficult to know how flexible the model needs to be in order to correctly model the true system. One solution to this problem is to use a model that is more flexible than necessary to make sure that the model is flexible enough. But how would that extra flexibility affect the uncertainty in the prediction? This is handled in the later part of the thesis where it is shown that the uncertainty in the prediction is bounded from below by the uncertainty in the prediction of the model with lowest flexibility required for representing true system accurately. In the literature, many other approaches to handle the uncertainty in predictions by NNs have been suggested, of which some are summarized in this work. Furthermore, a simulation and an experimental studies inspired by autonomous driving are conducted. In the simulation study, different sources of uncertainty are investigated, as well as how large the uncertainty in the predictions by NNs are in areas without training data. In the experimental study, the uncertainty in predictions done by different models are investigated. The results show that, compared to existing methods, the linearization method produces similar results for the uncertainty in predictions by NNs. An introduction video is available at https://youtu.be/O4ZcUTGXFN0<br>Inom forskning och utveckling har det har alltid varit centralt att skapa modeller av verkligheten. Dessa modeller har bland annat använts till att förutspå framtida händelser eller för att styra ett system till att bete sig som man önskar. Modellerna kan beskriva allt från hur friktionen hos ett bildäck påverkas av hur mycket hjulen glider till hur ett virus kan sprida sig i ett samhälle. I takt med att mer och mer data blir tillgänglig ökar potentialen för datadrivna black-box modeller. Dessa modeller är universella approximationer vilka ska kunna representera vilken godtycklig funktion som helst. Användningen av dessa modeller har haft stor framgång inom många områden men för att verkligen kunna etablera sig inom säkerhetskritiska områden såsom självkörande farkoster behövs en förståelse för osäkerhet i prediktionen från modellen. Neuronnät är ett exempel på en sådan black-box modell. I denna avhandling kommer olika sätt att tillförskaffa sig kunskap om osäkerhet i prediktionen av neuronnät undersökas. En metod som bygger på linjärisering av modellen för att tillförskaffa sig osäkerhet i prediktionen av neuronnätet kommer att presenteras. Denna metod är välbeprövad inom systemidentifiering och sensorfusion under antagandet att modellen är identifierbar. För modeller såsom neuronnät, vilka inte är identifierbara behövs det att det tas hänsyn till tvetydigheterna i modellen. En annan utmaning med datadrivna black-box modeller, är att veta om den valda modellmängden är tillräckligt generell för att kunna modellera det sanna systemet. En lösning på detta problem är att använda modeller som har mer flexibilitet än vad som behövs, det vill säga en överparameteriserad modell. Men hur påverkas osäkerheten i prediktionen av detta? Detta är något som undersöks i denna avhandling, vilken visar att osäkerheten i den överparameteriserad modellen kommer att vara begränsad underifrån av modellen med minst flexibilitet som ändå är tillräckligt generell för att modellera det sanna systemet. Som avslutning kommer dessa resultat att demonstreras i både en simuleringsstudie och en experimentstudie inspirerad av självkörande farkoster. Fokuset i simuleringsstudien är hur osäkerheten hos modellen är i områden med och utan tillgång till träningsdata medan experimentstudien fokuserar på jämförelsen mellan osäkerheten i olika typer av modeller.Resultaten från dessa studier visar att metoden som bygger på linjärisering ger liknande resultat för skattningen av osäkerheten i prediktionen av neuronnät, jämfört med existerande metoder.<br>iQdeep
50
Vu, Ky Minh. "System identification, control algorithms and control interval for the Box-Jenkins dynamic model structure." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1997. http://www.collectionscanada.ca/obj/s4/f2/dsk3/ftp04/nq25181.pdf.
Full text