No description available.

by Ka-lun Tung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1995. / Includes bibliographical references (leaves 168-[175]). / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Control with Uncertainties --- p.1 / Chapter 1.1.1 --- Adaptive Control --- p.2 / Chapter 1.1.2 --- H∞ Robust Control --- p.3 / Chapter 1.2 --- A Unified Framework: Adaptive Robust Control --- p.4 / Chapter 1.3 --- System Identification for Robust Control --- p.6 / Chapter 1.3.1 --- Choice of input signal --- p.7 / Chapter 1.4 --- Objectives and Contributions --- p.8 / Chapter 1.5 --- Thesis Outline --- p.9 / Chapter 2 --- Background on Robust Control --- p.11 / Chapter 2.1 --- Notation and Terminology --- p.12 / Chapter 2.1.1 --- Notation --- p.12 / Chapter 2.1.2 --- Linear System Terminology --- p.13 / Chapter 2.1.3 --- Norms --- p.15 / Chapter 2.1.4 --- More Terminology: A Standard Feedback Configuration --- p.17 / Chapter 2.2 --- Norms and Power for Signals and Systems --- p.18 / Chapter 2.3 --- Plant Uncertainty Model --- p.20 / Chapter 2.3.1 --- Multiplicative Unstructured Uncertainty --- p.21 / Chapter 2.3.2 --- Additive Unstructured Uncertainty --- p.22 / Chapter 2.3.3 --- Structured Uncertainty --- p.23 / Chapter 2.4 --- Motivation for H∞ Control Design --- p.23 / Chapter 2.4.1 --- Robust stabilization: Multiplicative Uncertainty and Weight- ing function W3 --- p.24 / Chapter 2.4.2 --- Robust stabilization: Additive Uncertainty and Weighting function W2 --- p.25 / Chapter 2.4.3 --- Tracking Problem --- p.26 / Chapter 2.4.4 --- Disturbance Rejection (or Sensitivity Minimization) --- p.27 / Chapter 2.5 --- The Robust Control Problem Statement --- p.28 / Chapter 2.5.1 --- The Mixed-Sensitivity Approach --- p.29 / Chapter 2.6 --- An Augmented Generalized Plant --- p.30 / Chapter 2.6.1 --- The Augmented Plant --- p.30 / Chapter 2.6.2 --- Adaptation of Augmented Plant to Sensitivity Minimiza- tion Problem --- p.32 / Chapter 2.6.3 --- Adaptation of Augmented Plant to Mixed-Sensitivity Prob- lem --- p.33 / Chapter 2.7 --- Using MATLAB Robust Control Toolbox --- p.34 / Chapter 3 --- Statistical Plant Set Estimation for Robust Control --- p.36 / Chapter 3.1 --- An Overview --- p.37 / Chapter 3.2 --- The Schroeder-phased Input Design --- p.39 / Chapter 3.3 --- The Statistical Additive Uncertainty Bounds --- p.40 / Chapter 3.4 --- Additive Uncertainty Characterization --- p.45 / Chapter 3.4.1 --- "Linear Programming Spectral Overbounding and Factor- ization Algorithm (LPSOF) [20,21]" --- p.45 / Chapter 4 --- Basic System Identification and Model Reduction Algorithms --- p.48 / Chapter 4.1 --- The Eigensystem Realization Algorithm --- p.49 / Chapter 4.1.1 --- Basic Algorithm --- p.49 / Chapter 4.1.2 --- Estimating Markov Parameters from Input/Output data: Observer/Kalman Filter Identification (OKID) --- p.51 / Chapter 4.2 --- The Frequency-Domain Identification via 2-norm Minimization --- p.54 / Chapter 4.3 --- Balanced Realization and Truncation --- p.55 / Chapter 4.4 --- Frequency Weighted Balanced Truncation --- p.56 / Chapter 5 --- Plant Model Reduction and Robust Control Design --- p.59 / Chapter 5.1 --- Problem Formulation --- p.59 / Chapter 5.2 --- Iterative Reweighting Scheme --- p.60 / Chapter 5.2.1 --- Rationale Behind the Scheme --- p.62 / Chapter 5.3 --- Integrated Model Reduction/ Robust Control Design with Iter- ated Reweighting --- p.63 / Chapter 5.4 --- A Design Example --- p.64 / Chapter 5.4.1 --- The Plant and Specification --- p.64 / Chapter 5.4.2 --- First Iteration --- p.65 / Chapter 5.4.3 --- Second Iteration --- p.67 / Chapter 5.5 --- Approximate Fractional Frequency Weighting --- p.69 / Chapter 5.5.1 --- Summary of Past Results --- p.69 / Chapter 5.5.2 --- Approximate Fractional Frequency Weighting Approach [40] --- p.70 / Chapter 5.5.3 --- Simulation Results --- p.71 / Chapter 5.6 --- Integrated System Identification/Control Design with Iterative Reweighting Scheme --- p.74 / Chapter 6 --- Controller Reduction and Robust Control Design --- p.82 / Chapter 6.1 --- Motivation for Controller Reduction --- p.83 / Chapter 6.2 --- Choice of Frequency Weightings for Controller Reduction --- p.84 / Chapter 6.2.1 --- Stability Margin Considerations --- p.84 / Chapter 6.2.2 --- Closed-Loop Transfer Function Considerations --- p.85 / Chapter 6.2.3 --- A New Way to Determine Frequency Weighting --- p.86 / Chapter 6.3 --- A Scheme for Iterative Frequency Weighted Controller Reduction (IFWCR) --- p.87 / Chapter 7 --- A Comparative Design Example --- p.90 / Chapter 7.1 --- Plant Model Reduction Approach --- p.90 / Chapter 7.2 --- Weighted Controller Reduction Approach --- p.94 / Chapter 7.2.1 --- A Full Order Controller --- p.94 / Chapter 7.2.2 --- Weighted Controller Reduction with Stability Considera- tions --- p.94 / Chapter 7.2.3 --- Iterative Weighted Controller Reduction --- p.96 / Chapter 7.3 --- Summary of Results --- p.101 / Chapter 7.4 --- Discussions of Results --- p.101 / Chapter 8 --- A Comparative Example on a Benchmark problem --- p.105 / Chapter 8.1 --- The Benchmark plant [54] --- p.106 / Chapter 8.1.1 --- Benchmark Format and Design Information --- p.106 / Chapter 8.1.2 --- Control Design Specifications --- p.107 / Chapter 8.2 --- Selection of Performance Weighting function --- p.108 / Chapter 8.2.1 --- Reciprocal Principle --- p.109 / Chapter 8.2.2 --- Selection of W1 --- p.110 / Chapter 8.2.3 --- Selection of W2 --- p.110 / Chapter 8.3 --- System Identification by ERA --- p.112 / Chapter 8.4 --- System Identification by Curve Fitting --- p.114 / Chapter 8.4.1 --- Spectral Estimate --- p.114 / Chapter 8.4.2 --- Curve Fitting Results --- p.114 / Chapter 8.5 --- Robust Control Design --- p.115 / Chapter 8.5.1 --- The selection of W1 weighting function --- p.115 / Chapter 8.5.2 --- Summary of Design Results --- p.116 / Chapter 8.6 --- Stress Level 1 --- p.117 / Chapter 8.6.1 --- System Identification Results --- p.117 / Chapter 8.6.2 --- Design Results --- p.119 / Chapter 8.6.3 --- Step Response --- p.121 / Chapter 8.7 --- Stress Level 2 --- p.124 / Chapter 8.7.1 --- System Identification Results --- p.124 / Chapter 8.7.2 --- Step Response --- p.125 / Chapter 8.8 --- Stress Level 3 --- p.128 / Chapter 8.8.1 --- System Identification Results --- p.128 / Chapter 8.8.2 --- Step Response --- p.129 / Chapter 8.9 --- Comparisons with Other Designs --- p.132 / Chapter 9 --- Conclusions and Recommendations for Further Research --- p.133 / Chapter 9.1 --- Conclusions --- p.133 / Chapter 9.2 --- Recommendations for Further Research --- p.135 / Chapter A --- Design Results of Stress Levels 2 and3 --- p.137 / Chapter A.1 --- Stress Level 2 --- p.137 / Chapter A.2 --- Stress Level 3 --- p.140 / Chapter B --- Step Responses with Reduced Order Controller --- p.142 / Chapter C --- Summary of Results of Other Groups on the Benchmark Prob- lem --- p.145 / Chapter C.1 --- Indirect and implicit adaptive predictive control [45] --- p.146 / Chapter C.2 --- H∞ Robust Control [51] --- p.150 / Chapter C.3 --- Robust Stability Degree Assignment [53] --- p.152 / Chapter C.4 --- Model Reference Adaptive Control [46] --- p.154 / Chapter C.5 --- Robust Pole Placement using ACSYDE (Automatic Control Sys- tem Design) [47] --- p.156 / Chapter C.6 --- Adaptive PI Control [48] --- p.157 / Chapter C.7 --- Adaptive Control with supervision [49] --- p.160 / Chapter C.8 --- Partial State Model Reference (PSRM) Control [50] --- p.162 / Chapter C.9 --- Contstrainted Receding Horizon Predictive Control (CRHPC) [52] --- p.165 / Bibliography --- p.168

Digitized by Kansas Correctional Industries

The next generation of intelligent vehicles will evolve from being able to drive autonomously to ones that communicate with other vehicles and execute joint behaviors. Before allowing these vehicles on public roads, we must guarantee that they will not cause accidents. We will apply formal methods to ensure the degree of safety that cannot be assured with simulation or closed-track testing. However, there are challenges that need to be addressed when applying formal verification techniques to cooperative driving systems. This thesis focuses on the techniques that address the following challenges: 1. Automotive applications interact with the physical world in different ways; 2. Cooperative driving systems are time-critical; 3. The problem of state explosion when we apply formal verification to systems with more participants. First, we describe the multiple stack architecture. It combines several stacks, each of which addresses a particular way of interaction with the physical world. The layered structure in each stack makes it possible for engineers to implement cooperative driving applications without being bogged down by the details of low-level devices. Having functions arranged in a layered fashion helps us divide the verification of the whole system into smaller subproblems of independent module verification. Secondly, we present a framework for modeling the protocol systems that uses GPS clocks for synchronization. We introduce the timing stack, which separates a process into two parts: the part modeled as an finite-state machine that controls state transitions and messages exchanges, and the part that determines the exact moment that a timed event should occur. The availability of accurate clocks at different locations allows processes to execute actions simultaneously, reducing interleaving that often arises in systems that use multiple timers to control timed events. With accurate clocks, we create a lock protocol that resolves conflicting merge requests for driver-assisted merging. Thirdly, we introduce stratified probabilistic verification that mitigates state explosion. It greatly improves the probability bound obtained in the original probabilistic verification algorithm. Unlike most techniques that aim at reducing state space, it is a directed state traversal, prioritizing the states that are more likely to be encountered during system execution. When state traversal stops upon depleting the memory, the unexplored states are the ones that are less likely to be reached. We construct a linear program whose solution is the upper bound for the probability of reaching those unexplored states. The stratified algorithm is particularly useful when considering a protocol system that depends on several imperfect components that may fail with small but hard-to-quantify probabilities. In that case, we adopt a compositional approach to verify a collection of components, assuming that the components have inexact probability guarantees. Finally, we present our design of driver-assisted merging. Its design is reasonably simplified by using the multiple stack architecture and GPS clocks. We use a stratified algorithm to show that merging system fails less than once every 5 × 10¹³ merge attempts.

Cooperative control of networked systems (or multi-agent systems) has attracted much attention during the past few years. But most of the existing results focus on first order and second order leaderless consensus problems with linear dynamics. The second part of this dissertation solves a higher-order synchronization problem for cooperative nonlinear systems with an active leader. The communication network considered is a weighted directed graph with fixed topology. Each agent is modeled by a higher-order nonlinear system with the nonlinear dynamics unknown. External unknown disturbances perturb each agent. The leader agent is modeled as a higher-order non-autonomous nonlinear system. It acts as a command generator and can only give commands to a small portion of the networked group. A robust adaptive neural network controller is designed for each agent. Neural network learning algorithms are given such that all nodes ultimately synchronize to the leader node with a small residual error. Moreover, these controllers are totally distributed in the sense that each controller only requires its own information and its neighbors' information. / Receding horizon control (RHC), also called model predictive control (MPC), is a suboptimal control scheme over an infinite horizon that is determined by solving a finite horizon open-loop optimal control problem repeatedly. It has widespread applications in industry. Reinforcement learning (RL) is a computational intelligence method in which an optimal control policy is learned over time by evaluating the performance of suboptimal control policies. In this dissertation it is shown that reinforcement learning techniques can significantly improve the behavior of RHC. Specifically, RL methods are used to add a learning feature to RHC. It is shown that keeping track of the value learned at the previous iteration and using it as the new terminal cost for RHC can overcome traditional strong requirements for RHC stability, such as that the terminal cost be a control Lyapunov function, or that the horizon length be greater than some bound. We propose improved RHC algorithms, called updated terminal cost receding horizon control (UTC-RHC), first in the framework of discrete-time linear systems and then in the framework of continuous-time linear systems. For both cases, we show the uniform exponential stability of the closed-loop system can be guaranteed under very mild conditions. Moreover, unlike RHC, the UTC-RHC control gain approaches the optimal policy associated with the infinite horizon optimal control problem. To show these properties, non-standard Lyapunov functions are introduced for both discrete-time case and continuous-time case. / Two topics of modern control are investigated in this dissertation, namely receding horizon control (RHC) and cooperative control of networked systems. We apply learning techniques to these two topics. Specifically, we incorporate the reinforcement learning concept into the standard receding horizon control, yielding a new RHC algorithm, and relax the stability constraints required for standard RHC. For the second topic, we apply neural adaptive control in synchronization of the networked nonlinear systems and propose distributed robust adaptive controllers such that all nodes synchronize to a leader node. / Zhang, Hongwei. / Adviser: Jie Huang. / Source: Dissertation Abstracts International, Volume: 72-04, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2010. / Includes bibliographical references (leaves 99-105). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

<p>Scania är en välkänd tillverkare av tunga lastfordon och bussar. De profilerar sig som det presigefyllda valet med hög teknologinivå. För att kunna leda utvecklingen av nya funktioner och motorer måste många tester av alla system i dessa fordon göras. Till detta används provceller speciellt ordnade för specifika lastbilskomponenter och många av dessa komponenter behöver en yttre kylning under drift varför flera av provcellerna har reglerade kylsystem. Vid uppgradering av kylsystem eller nybyggnation av provceller med kylsystem finns en önskan att kunna simulera och göra tester av kylsystemet innan det faktiska kylsystemet finns på plats, för att säkerställa prestandan innan någon investering görs. Detta examensarbete går ut på att genom att skapa vissa basmodeller för komponenter i kylsystem, möjliggöra simulering av godtyckliga kylsystem för utvärdering av olika reglerstrategier, designer och deras prestanda. Vid framtagning av modellerna användes mestadels fysikalisk modellering men även rent praktiska modeller för att fylla en funktion existerar. Det resulterande modellbiblioteket klarar av att visa dynamiska effekter mycket bra men dess statiska träffsäkerhet är inte lika exakt. Dynamiken är dock det viktiga och svårt att få fram på andra sätt, varför modellbiblioteket kan vara användbart vid bedömningar om reglerstrategier och prestanda.</p> / <p>Scania is a wellknown manufacturer of heavy transport vehicles such as trucks and buses. Their profile is to be the prestigeous choice with a high level of technology. To maintain the leading position in development of new functionality and engines, substantial testing is nessecary. This is accomplished by using testbeds specialized for different components and purposes. Many of these components need external cooling during operation which is why a cooling system is present in a large number of testbeds. When new testbeds are to be constructed or an old cooling system is to be redesigned, there is a desire to be able to simulate and run tests of the cooling system before the actual cooling system is in place, to make sure performance is at a high enough level, before any investment is made. The task in this master thesis is to construct, in a matlab and simulink environment, a set of base models for cooling system components. With this set, arbitrary cooling systems can be constructed and simulated to evaluate different designes, control strategies and performance. Physical modelling was the most common method when constucting the base models though some models have a pure practical function. The resulting model library is able to, when put together to a complete cooling system, show dynamic behaviour correctly but static accuracy is a bit off. When judging a new control strategy, dynamic behaviour is the most important aspect, and also the most difficult to get elsewhere, which is why the model library can still be very useful.</p>

<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>

<p>A cars powertrain consists of everything that is needed for its propulsion. The components in the driveline that transfer the power from the engine to the driven wheels are not absolutely stiff, hence they will wind up due to the torque and act as torsion springs. If you suddenly demand a bigger torque by stepping on the accelerator pedal, a so called tip in manoeuvre, and that torque is acquired from the engine as quickly as possible, the driveline will not be able to transfer that fast torque change due to its weakness and as a result it will start to oscillate. These oscillations will be transferred to the driven wheels and make the car to accelerate jerkily which will be experienced as uncomfortable by the passengers. Furthermore, there is a backlash in the driveline that will make the weakness in the driveline even more excited than it should have been if the backlash did not existed.</p><p>To avoid too big problems with these oscillations there is a control system that controls the demanded torque. This control system is today an open loop control system, i.e. a filtering of the demanded torque. As the cars computer power is increasing steadily there is an interest of investigating if it is possible to get a higher performance control system by using a more advanced, closed loop, model based control system.</p><p>In this thesis such a control system is developed. First a model of the system is constructed; this model is used to design an observer that estimates the non measurable states in the driveline. The results show that this observer estimates these states fine on the basis of the available signals. The observer is the base for the driveline control system and simulations show that this control system is a considerable improvement compared to the control system used today.</p> / <p>En bils drivlina består av allt som behövs för dess framdrift. De komponenter i drivlinan som överför kraften från motorn till hjulen är inte absolut stela utan dessa elastiska kroppar kommer att deformeras då de utsätts för ett moment och fungera som torsionsfjädrar. Om man snabbt begär ett större moment genom att trampa på gasen, en så kallad tip-in manöver, och detta begärda moment läggs ut så snabbt som möjligt från motorn kommer inte de veka axlarna ”hinna med” vilket medför att axlarna kommer att börja oscillera. Dessa oscillationer kommer att överföras till de drivande hjulen och göra att bilen accelererar ryckigt vilket upplevs som obehagligt av passagerarna i bilen. Detsamma gäller om man plötsligt släpper gasen, en så kallad tip-out manöver. Det finns dessutom ett glapp i drivlinan som gör att vekheten i axeln exciteras ännu mer än vad den skulle ha gjort om glappet inte hade existerat.</p><p>För att undvika alltför stora problem med dessa oscillationer har man en reglering som gör att man inte lägger på hela det begärda momentet på en gång. I dag fungerar denna reglering till största delen genom öppen styrning, det vill säga genom en filtrering av det begärda momentet. I takt med att datorkraften hela tiden ökar i bilar är man nu intresserad av att utreda i fall man kan lösa detta reglerproblem på ett bättre sätt genom en mer avancerad, sluten, modellbaserad reglering. Även i den nya regleringen vill man enbart använda sig av redan befintliga sensorer i bilen då det skulle bli för dyrt att sätta in någon extra sensor enbart för att få en bättre drivlinereglering.</p><p>I denna uppsats utvecklas en sådan reglering. Först konstrueras en modell av drivlinan som används till att skapa en observatör som skattar de icke mätbara tillstånden i drivlinan. Resultaten visar på att denna observatör skattar de icke mätbara tillstånden på ett bra sätt utifrån de tillgängliga mätsignalerna. Denna observatör används sedan som grund för en reglering av drivlinan och i simuleringar visar sig denna reglering vara markant bättre än den som används idag.</p>

<p>The aim of this thesis is to develop the knowledge about nonlinear and/or adaptive solid fuel boiler control at Vattenfall Utveckling AB. The aim is also to make a study of implemented and published control strategies.</p><p>A solid fuel boiler is a large-scale heat (and power) generating plant. The Idbäcken boiler studied in this work, is a one hundred MW furnace mainly fired with wood chips. The control system consists of several linear PID controllers working together, and the furnace is a nonlinear system. That, and the fact that the fuel-flow is not monitored, are the main reasons for the control problems. The system fluctuates periodically and the CO outlets sometimes rise high above the permitted level.</p><p>There is little work done in the area of advanced boiler control, but some interesting approaches are described in scientific articles. MPC (Model Predictive Control), nonlinear system identification using ANN (Artificial Neural Network), fuzzy logic, Hµ loop shaping and MIMO (Multiple Input Multiple Output) PID tuning methods have been tested with good results.</p><p>Both linear and nonlinear system identification is performed in the thesis. The linear models are able to explain about forty percent of the system behavior and the nonlinear models explain about sixty to eighty percent. The main result is that nonlinear models improve the performance and that there are considerable disturbances complicating the identification. Another identification issue was the feedback during the data collection.</p>

<p>I fumingverket på Rönnskärsverken utvinns zinkklinker ur slaggen från elugnen. En annan råvara är stålverksstoft. I fumingugnen omvandlas smältans zink- och blyinnehåll till metallånga som oxideras till ett stoft. Stoftet renas i en klinkerugn. Slutprodukten, zinkklinker, som består av 70 - 75 % zink, exporteras till zinksmältverket Norzink i Norge. Klinkerugnen är ett väldigt långsamt system med stegsvarstider i storleksordningen en timme vilket gör den svårstyrd och det resulterar i störningar och stilleståndstid med låg produktion och låg kvalitet på klinkern. För att lösa detta problem testas automatisk reglering i detta arbete. Två processmodeller tas fram för simulering och reglerdesign och tre regulatorer har utvecklats i simulering. Två av dessa testas på den faktiska processen. Vidare har ett ramverk för snabb utveckling och testning av regulatorer utvecklats. En C++-klass för kommunikation via DDE-gränssnittet mellan regulator och operatörsgränssnittet har också konstruerats.</p> / <p>In the fuming plant at Rönnskärsverken smelter, zinc clinker is xtracted from slags and steel mill dust. In the fuming furnace, zinc and lead are vapourised by coal injection. The reoxidised metal dust is further refined at the clinker plant to obtain a product that is low in halogenes. Zinc clinker, which contains approximately 70 - 75 % zinc, is exported to the Norzink zinc smelter in Norway. The refinement takes place in an industrial kiln. The kiln is a very slow system and therefore difficult to control which results in disturbances and dead time. This causes low production rate and poor quality in the clinker. In order to cope with this, automatic control is tested in this thesis. Two process models have been built for simulation and control design and three controllers have been evaluated in simulation. Two of the developed controllers are tested on the actual process. A framework for fast controller prototyping has also been developed. A C++-class för communication using the DDE interface between controller and the operator user interface has also been implemented.</p>