<p>At excitatory synapses, NMDAR-mediated synaptic plasticity occurs in response to activity inputs by modifying synaptic strength. While comprehensive studies have been focused on the induction and expression mechanisms underlying synaptic plasticity, it is less clear whether and how synaptic plasticity itself can be subjected to regulations. The presence of "plasticity of plasticity", or meta-plasticity, has been proposed as an essential mechanism to ensure a proper working range of plasticity, which may also offer an additional layer of information storage capacity. However, it remains elusive whether and how meta-plasticity occurs at single synapses and what molecular substrates are locally utilized. Here, I develop systems allowing for sustained alterations of individual synaptic inputs. By implementing a history of inactivity at single synapses, I demonstrate that individual synaptic inputs control synaptic molecular composition homosynaptically, while allowing heterosynaptic integration along dendrites. Furthermore, I report that subunit-specific regulation of NMDARs at single synapses mediates a novel form of input-specific metaplasticity. Prolonged suppression of synaptic releases at single synapses enhances synaptic NMDAR-mediated currents and increases the number of functional NMDARs containing NR2B. Interestingly, synaptic NMDAR composition is adjusted by spontaneous glutamate release rather than evoked activity. I also demonstrate that inactivated synapses with more NMDARs containing NR2B acquire a lower induction threshold for long-term synaptic potentiation. Together, these results suggest that at single synapses, spontaneous release primes the synapse by modifying its synaptic state with specific molecular compositions, which in turn determine the synaptic gain in an input-specific manner.</p> / Dissertation

The utilization of micro-scale technologies is limited by the speed of their manufacture. Micro-milling is one particular technology used to manufacture micro-scale parts which could benefit extensively from an increase in throughput. Micro-milling involves a rotating cutter slightly thicker than a human hair removing material while spinning at speeds often over one hundred thousand revolutions per minute. An obvious solution to the throughput bottleneck is to move current micro-mills faster using existing technology; however, simply increasing the operational speed of existing micro-mills will lead to vibration and trajectory following problems. If a micro-mill cannot be positioned precisely, then part tolerances cannot be maintained. Thus any increase in throughput would be counterproductive in terms of overall performance. This dissertation presents techniques to improve the performance of micro-mills, as well as other flexible machines. Theses improvements are possible through the utilization of the vibration suppression scheme of input shaping. By thoughtfully altering the commands sent to flexible systems, their vibration can be significantly reduced. Input shaping was effectively applied to an existing micro-mill, which improved part tolerances and increased operational speed. However, at extremely high speeds, traditional input shaping is not effective at following complicated trajectories. Therefore, new input shaping techniques were developed specifically for trajectory tracking of extremely fast motions on micro-mills and other flexible systems. Often machines cannot achieve these high speeds while maintaining their accuracy because of the mechanical design of the machines themselves. If the mechanical design of micro-mills and other machines consider flexible and lightweight design alternatives that utilize input shaping for vibration suppression instead of stiff and heavy designs, then faster machine motion will be possible. By considering input shaped flexible systems as part of traditional mechanical design processes, these flexible solutions allow vast performance improvement. Specifically, embodiment design can be improved through consideration of input shaping performance requirements. Through these advancements, this dissertation improves the design, control, and performance of micro-mills and other flexible machines.

Command Generation has been shown to be a practical and effective control scheme for eliminating payload swing on industrial cranes. However, this technology has not been used to its full potential. One reason is that nonlinear crane dynamics degrade the performance of current command generators, making them challenging to use. A second reason is that few crane operators are aware of this technology. Therefore, this thesis strives to alleviate these problems through the completion of three major tasks. First, new command generation algorithms are developed that compensate for nonlinear crane dynamics. Two major sources of non-linear dynamics are targeted: nonlinear drive dynamics, and non-linear physical dynamics of tower cranes. Second, command generation are examined from an educational perspective; both in the classroom and in the working field. Third, three experimental crane devices were built to fulfill the two prior tasks.

The main contents of this thesis are improving a UWB LNA, and analyze the input-matching, the noise, and the gain. First we increase the width of the input transistor, and remove source-degeneration inductor. Those ways can increase the gain and reduce the noise of the circuit. In the input matching, we use a shunt capacitor, a series inductor, and the impedance of the transistor itself to achieve high frequency matching. The lower frequency matching is achieved by negative feedback resistor. The UWB LNA dissipates 10.14 mW power and achieves input return loss (S11) below -11.5 dB, output return loss (S22) below -11.9 dB, forward gain (S21) of 14.4¡Ó0.4 dB, reverse isolation (S12) below -26.7 dB, and noise figure (NF) of 2.6~3.5 dB over the 3.1~10.6 GHz band of interest. 1-dB compression point (P1dB) of -16.8 dBm and input third-order inter-modulation point (IIP3) of -8.1 dBm are achieved at 6.85 GHz.

In this paper, we investigate a highly efficient input method in Chinese. In the traditional Mandarin phonetic input method, users have to input the complete Mandarin phonetic symbol. The proposed new Chinese input method is which transforms the first Mandarin phonetic symbol sequence to character sequence. Users only have to input the first Mandarin phonetic symbol. Users input first Mandarin phonetic symbol and follow the input rule that spaces are inserted between the words. The system outputs the candidate character sequence hypotheses. Bigram model is used to describe the relation between words. We use the dynamic programing for decoding. We estimate the feasibility for our new Chinese input method and estimate the Stanford segmenter. In the experiment, we estimate the Standford Segmenter works on the simplified Chinese and Traditional Chinese firstly. We observe that the precision and recall on the simplified Chinese are 84.52% and 85.20% which is better than works on the Traditional Chinese 68.43% and 63.43%. And we estimate system efficiency based on language model that trained by WIKI corpus and ASBC corpus separately. The sentence and word accuracy for the ASBC corpus are 39.8% and 70.3%. And the word and character accuracy for WIKI corpus are 20.3% and 53.3%. Finally we estimate the number of candidate hypotheses. The research shows the 10 hypotheses and 20 hypotheses the sentence accuracy are closed.

The main contents of this thesis are improving a UWB LNA, and analyze the input-matching, the noise, and the gain. First we use the feedback of the input transistor , and it different from the traditional source-degeneration inductor.The design can increase the gain and reduce the noise of the circuit.The second stage CS architecture designed to improve the overall gain of the circuit. Output level to use the source follower with the device even when the output matching . In the input matching,we use a shunt inductor and the impedance of the transistor itself to achieve high frequency matching. The UWB LNA dissipates 16.8 mW power and achieves input return loss (S11) -9.3 to -10 dB, output return loss (S22) -16.83 to -13 dB, forward gain (S21) 13.8 to 11.6 dB, reverse isolation (S12) below -30 dB, and noise figure (NF) of 2.38~3.31 dB over the 6~10 GHz band of interest. 1-dB compression point (P1dB) of -12.5 dBm and input third-order inter-modulation point (IIP3) of -2.5 dBm are achieved at 6 GHz.

A method of controller design for multiple input multiple output (MIMO) system is needed that will not give the high order controllers of modern control theory but will be more systematic than the “ad hoc” method. The objective of this method of design for multiple input multiple output systems is to find a controller of fixed order with performance specifications taken into consideration. An inner approximation of the stabilizing set is found through the algorithm discussed in Keel and Bhattacharyya’s "Fixed order multivariable controller synthesis: A new algorithm." The set satisfying the performance is then approximated through one of two algorithms; a hybrid of two optimization algorithms or the grid algorithm found in Lampton’s "Reinforcement Learning of a Morphing Airfoil-Policy and Discrete Learning Analysis." The method is then applied to five models of four aircraft; Commander 700, X-29, X-38, and F-5A using controllers of first and second orders.

This dissertation investigates the dynamic economic structure transformation and its corresponding environmental consequences at the Austin-San Marcos Metropolitan Statistical Area (Austin MSA) from 1990 to 2008. Input-output (IO) analysis is the major methodology and environmental problems are defined as emissions of industrial point air pollutants. Both three-and seven-segment IO models of Austin MSA for the years of 1990, 1994, and 1999 are constructed. Direct and total pollution coefficients of six major pollutants are calculated, hypothe tical extraction measurement and structural decomposition analysis are implemented, and the quantity and pattern of pollutant emissions are simulated based on four major assumed development scenarios from 2000 to 2008. This study finds: 1) the digital economy has emerged in the Austin MSA during the 1990s, 2) the manufacturing process of Austin MSA tended to be more environmentally friendly, which supports the hypothesis of dematerialization and decarbonization, 3) consumption-driven and non-production segments related environmental problems becomes more significant in the emerging digital economy. This study predicts that industrial point air pollutant emissions will grow moderately from 2000 to 2008, assuming that the direct pollutant coefficients will change at the average rates of the 1990s and the final demand will grow at the half rates of the 1990s?? average. Pollution contribution from production segment will generally decrease and contribution from other segments such as ICT and Information will increase, however, emission contributions of the segments will vary in terms of pollutants as well as development scenarios. This study argues that the shift of the source and nature of environmental threats of in the digital economy mandates parallel reform of the current environmental policy. A new generation of policy should be cooperative rather than confrontational, integrated rather than fragmented, flexible rather than rigid. It should also facilitate innovative management initiatives to achieve sustainability. More fundamentally, it is expected to deal with environmental impacts of intangible information flows (bits) which are possibly more essential than flows of tangible goods and services (atoms) in the context of the digital economy and the information age.

A novel algorithm for making the combination of outputs in the output zero direction of the plant always equal to zero was formulated. Using this algorithm and the result of MacFarlane and Karcanias, a fault detection scheme was proposed which utilizes the directional property of the multivariable linear system. The fault detection scheme is applicable to linear multivariable systems. Results were obtained for both continuous and discrete linear multivariable systems. A quadruple tank system was used to illustrate the results. The results were further verified by the steady state analysis of the plant.

Based on the Lyapunov stability theorem, a design methodology of adaptive block backstepping controller is proposed in this thesis for a class of multi-input systems with mismatched perturbations to solve regulation problems. According to the number of block (m) in the plant to be controlled, m-1 virtual input controllers are designed from the first block to the (m-1)th block. Then the proposed robust controller is designed from the last block. Adaptive mechanisms are employed in each of the virtual input controllers as well as the robust controller, so that the knowledge of the least upper bounds of mismatched perturbations is not required. The resultant control system can achieve asymptotic stability. Finally, a numerical example and a practical example are given for demonstrating the feasibility of the proposed control scheme.