The Assessment System implemented in this research has three characteristics. First ItemBank standardization¡Afollow and import IMS QTI specifications. Then, we can reuse ItemBank resource in other platforms which follow IMS QTI specification. Second ItemBank maintainable¡Athrid feedback information for teacher and student . Both implemented in feedback design through the assessment analysis agent.

A review of adaptive decorrelating detector techniques for direct-sequence code division multiple access (CDMA) signals is given. The goal is to improve CDMA system performance and capacity by reducing interference between users. The techniques considered are implementations of multiuser receivers, for which background material is given. Adaptive algorithms improve the feasibility of such receivers. An adaptive multiuser receiver for CDMA system has been proposed to increase system capacity. The adaptive decorrelator can be used to eliminate interference from known interferers, though it is prone to noise enhancement. However the receiver is basically designed for synchronous CDMA over AWGN channels. In order to confirm the robustness of the receiver for the asynchronous cases, experimental evaluations are displayed when the relative user delays are small compared to the symbol duration and when the channel is Rayleigh multipath fading as in micro-cellular scenarios. In addition to the efficient implementation of the decorrelating detector of [1], the receiver also can be adapted to incorporate decision feedback. Successive interference cancellation techniques reduce interference by cancellation of one detected signal from another. And an efficient incorporation of decorrelator with RAKE and (DF) decision feedback receiver for frequency-selective Rayleigh fading multipath channels is also proposed. Performance evaluation of the detector via computer simulation scenarios is conducted to substantiate it's potential for real-time operation.

This work investigated the characteristics of transverse modes and polarization of a VCSEL and attempted to control its transverse mode with optical feedback. The fundamental studies indicated that this solitary VCSEL's output consisted of two mutually orthogonal linear polarizations with different threshold currents and different transverse mode distributions. Furthermore, the temperature dependence of the transverse modes of the VCSEL revealed a better stability than that of edge-emitted lasers. In the mode controlling experiments, an iris was employed to select a specific transverse mode for feedback, resulting in an enhancement in the power of the feedback mode. A significant improvement was accomplished in the feedback of the fundamental mode of the major polarization component: the laser oscillated in a single mode with a side mode suppression ratio greater than 19.8 dB and with a depression of the relative intensity noise more than 20 dB/Hz. Nevertheless, a week feedback in the minor polarization component would stimulate more competition among transverse modes and degraded the performance of the VCSEL. Finally, no effect was observed in the orthogonal-polarization optical feedback experiment due to the excessively small of feedback ratio.

This thesis considers the design of cascaded SISO control systems using the H∞ and QFT methodologies. In the first part of the thesis the actual advantages offered by Single Input Single Output (SISO) cascade loop structures are studied. In Quantitative Feedback Theory(QFT) it is emphasized that the use of cascaded loops is primarily for the reduction of bandwidth of the controllers. This in turn helps in considerable reduction of the adverse effects of high frequency noise. The question that arises then is whether or not there are any substantial benefits to be gained by cascade loop design in the low frequencies. It is shown using QFT methodology that there aren’t any advantages gained in the low frequencies with the use of cascaded design. In effect it is concluded that if the design is properly executed a single loop controller closed from the output to the input will be sufficient to meet the typical performance specifications. This is shown using an example where the mold level of a continuous casting process is to be controlled. The plant being used has considerable uncertainty so that features of robust control can be highlighted. In the second part the Robust Outer Loop bounds were generated analytically and examined for certain properties. It was compared to the bounds generated by already existing algorithms. In the third part the inner outer QFT design was modified with the inner loop being designed using H∞ with the concept of sensitivity shaping. This design was very similar to the pure QFT design with the added advantage of having some automation. In the fourth part the H∞ methodology was used to design a two loop control structure. The idea was to compare this design to the QFT design. It was seen that H∞ generated redundant controllers and pre filters.

The use of optical feedback for improving response linearity of electrooptic light modulators has been investigated. The modulator is configured as a straight channel waveguide flanked by electrodes in a lithium niobate (LiNbO3) substrate. Light is coupled into the waveguide in both TE and TM polarizations, and a voltage applied across electrodes causes a relative phase shift between two polarization components. An output analyzer converts the phase modulation to intensity modulation. Optical feedback of light in both polarization modes results from reflection of light at the polished edges of the substrate. Channel waveguides supporting a single guided mode for TE and TM polarizations were fabricated in x-cut LiNbO3 substrates using titanium-indiffusion technique. The waveguides and modulators were characterized at a wavelength of 1.55 ??m using a distributed feedback laser. The modulators were driven with a sinusoidal voltage waveform. To minimize harmonics of the modulating frequency in the intensity output, the magnitude of the optical feedback and the substrate temperature were adjusted. The feedback level was altered by applying refractive index-matching liquid to one or both ends of the waveguide at the edges of the crystal. It was found that a high degree of response linearity in the presence of feedback was achievable at certain substrate temperatures. The spurious-free dynamic range (SFDR) relative to the noise floor was measured at different feedback levels and substrate temperatures in an effort to maximize the modulator response linearity. An SFDR of 68.04 dB, limited by third-order nonlinearity, was achieved by applying index-matching fluid to the input end of the substrate. This compares with an SFDR of 64.84 dB limited by second-order nonlinearity when index-matching fluid was applied at both ends of the substrate. By changing the temperature of the same substrate to adjust the phase shifts experienced by TE and TM polarizations, the SFDR with index-matching fluid at the input end increased to 71.83 dB, limited by third-order nonlinearity. In tests at constant modulation depth, an improvement of as much as 9.6 dB in SFDR vs. the theoretical value for an interferometric modulator without feedback was achieved.

Autoland controllers are prevalent for both large and small/micro unmanned aerial vehicles, but very few are available for medium sized unmanned aerial vehicles. These vehicles tend to have limited sensors and instrumentation, yet must possess good performance in the presence of modeling uncertainties, and exogenous inputs such as turbulence. Quantitative Feedback Theory is an attractive control methodology for this application, since it provides good performance and robustness for systems with structured model uncertainties. It has been successfully applied to many aircraft problems, but not to automatic landing, and only inner-loop synthesis has been presented in the literature. This paper describes the synthesis and development of an automatic landing controller for medium size unmanned aerial vehicles, using discrete Quantitative Feedback Theory. Controllers for the localizer, glideslope tracker, and automatic flare are developed, with a focus on the outer-loops synthesis. Linear, non real-time six degree-of-freedom Monte Carlo simulation is used to compare the Quantitative Feedback Theory controller to a baseline Proportional-Integral controller in several still air and turbulent landing scenarios. The Quantitative Feedback Theory controller provides performance similar to the Proportional-Integral controller in still and in turbulent air. Both controllers show similar robustness to turbulence, but the Quantitative Feedback Theory controller provides significantly better robustness to model uncertainties in turbulent air as well as to sensor characteristics in turbulence. Based on the results of the paper, the QFT controller is a promising candidate for an autoland controller.

Microsystem and integrated circuitry components are mostly manufactured using semiconductor technologies. Fabrication using high strength metals, for demanding aerospace, mechanical, or biomedical applications, requires novel technologies which are different from those for silicon. A promising mass production method for micro/meso scale components is electrochemical micromachining. The complex system, however, requires high precision mechanical fixtures and sophisticated instrumentation for proper process control. This study presents an electrochemical micromachining system with a closed-loop feedback control programmed using a conditional binary logic approach. The closed-loop control is realized using electrical current as the dynamic feedback signal. The control system improves material removal rate by 250% through optimizing inter electrode gap and provides robust automation reducing machining variation by 88%. The new system evokes production of higher quality microcomponents. Workpiece damage is reduced by 97% and increased feature sharpness is observed.

Based on the theory of System dynamics and by using the methodology of a case study, the researcher explores the factors influencing the shortened product lead time and the stability of the product quality. In Taiwan, the industry of after-market car accessory has been facing the pressure resulted from our competition with mainland China to provide low-price products. Such competition has intensified since frequent business communications and encounters between Taiwan and China take place. In order to thrive in such competition, not only does the industry in Taiwan need to increase its productivity, but it also has to find means to increase the stability of the product quality and to shorten the product lead time. Analyzing it based on the process of a production, one can find that product lead time is influenced by the following three factors: the time required to design the product; to purchase the raw materials; to produce the product. When purchasing raw materials, one has to consider both the materials used in product developing and as well as the ones required to mass production. In addition, the time required to acquire raw materials also depends on the reliability of the up-stream supplier and our relations with them. Last but not least, the time required to acquire raw materials also depend upon how professional the purchaser designated buyer in the company is. The time required to design the product is affected by the R/D engineer¡¦s professionalism and his or her familiarity with the product. Consequently, the time required for designing phase is affected by the possible turn-over of the company¡¦s engineers. Further, the design and equipment of the production line and workers¡¦ familiarity with the production process are all determine factors as to the time required to complete the products. The quality of a product is determined by the quality of the raw materials and the process of mass production. The quality of the out sourced raw materials is affected by the attitude and promptness of the purchasing department of the company. Furthermore, the standard operation process of the production line and the design of all concerned fixtures also affect the product quality during the mass production. By using the Strategy Maps and Balanced Scorecard, one found that, in the case of Company S, the instability of the quality of products and the delayed product lead time were caused by the lack of competency of company buyers, poor management of the supply chains, and ineffectively executed quality control process. In order to improve the situation for Company S, one needs to consider the interactions between all production factors based on the perspectives of the System Thinking. Subsequently, based on analyzing the case of Company S, the current study concludes the following points: 1. Using traditional business management models, the decision-makers cannot detect the dynamics of processes in the company. 2. By only using the Casual Feedback, one cannot easily detect the core of the problem in the company. One also needs to simulate the situation in order to foresee the result. 3. One should use a system dynamic way of thinking to analyze the operations of a corporation in order to foresee potential problems before they arise. 4. The researcher will recommend Company S to establish its own Strategy Mapping, so it can effectively evaluate its policies and strategies. Key Word: System dynamics, System Thinking, Strategy Maps, The Balanced Scorecard, Casual Feedback,

Agents in financial markets must learn about underlying variables and hence find some patterns of parameter. Learning schemes across agents may be different since agents¡¦ learning procedure depends on their computational technique and their beliefs. This dissertation outlines how agents learn from the actions of others or observe relative information when making trading decisions. If prior or signal follow a normal distribution, the law of standard Gaussian updating can be applied to represent the process of learning by agents. Accordingly, this dissertation use learning and global game technique to solve for the condition of equilibrium, to discuss the problem of eliminating multiplicity and to analyze three topics as follow. First, how transparency in financial market influences equilibrium condition, ad hoc multiplicity, is rarely to explore. The first part of this dissertation focuses on the market-clearing condition, based on a noisy rational expectations equilibrium and using the global game technique with learning to proposes a policy effect and the variant of equilibrium in financial market. The after which conclusions can be drawn if transparent policy in financial market is implemented. First, all else equal, regardless of the precision of private signals of uninformed traders, financial market transparency cannot prevent from the vanishing of multiplicity under exogenous dividend return. Next, when dividend return is endogenous, the unique equilibrium condition is partly determined by the precision of private signals from uninformed traders rather than perfectly by a policy of market transparency. Second, politicians and some scholars have advocated that emerging market countries avoid financial instability and reduce multiplicity by restricting capital mobility. Moreover, in the highly sensitized financial markets, information released by the authority is often seen so as to reply rapidly market variety and to correct and remove irrational expectations. The secondary direction of this dissertation employs direct capital mobility controls and informational releases to discuss and compare the variant of equilibrium. A notable finding is that, although direct capital mobility controls and informational releases may successfully reduce capital outflows, direct restrictions on capital mobility are more likely to encourage speculative attack under optimistic signals and are thus more likely to cause multiple equilibriums, ceteris paribus. From a policy perspective, under a signal is optimistic, maintaining uniqueness is more likely when information release is aggressive than when information release is passive or unrestricted on capital mobility, ceteris paribus. Finally, recent studies about price or crises focus on the different directions or targets of learning. However, the problems that learning may generate endogenous feedback and endogenous coordination incentive and affect market performance are rarely discussed. Hence, the third topic of this dissertation will spotlight learning about aggregating information and observation of market and explain the effect of either coordination incentive or feedback on excess volatility. As a result, feedback effects may result from coordination incentives among agents and more coordination incentives among agents increases the feedback effect. Next, when financial markets are highly liquid or when financial shocks are severe, coordination incentives among agents decline. Excess asset price volatility decreases with either feedback effects or coordination incentives.

Thesis (Ph. D.)--University of Washington, 2001. / Vita. Includes bibliographical references (leaves 131-139).