Two robust control strategies are proposed in this dissertation for a class of multi-input multi-output dynamic systems with matched and mismatched perturbations. First of all, a novel design methodology of switching variables is proposed for solving the regulation problems. A serial state transformations are needed in order to design pseudo feedback gains and adaptive mechanisms. By utilizing the pseudo control input gain embedded in each of the switching variable, the proposed controller can not only suppress the mismatched perturbations when the controlled systems are in the sliding mode, but also attain locally asymptotic stability. The design of a robust output tracking controller is presented next for solving the tracking problems. Without utilizing the information of state variable, the proposed output feedback tracking controllers are capable of driving the state tracking errors into a small bounded region whose size can be adjusted through the designed parameters, and guarantee the stability of controlled systems. These two robust control schemes are designed by means of the variable structure control technique with sliding mode and Lyapunov stability theorem. Each controller contains three parts. The first part is for eliminating measurable feedback signals. The second part is used for adjusting the convergent rate of state variables (or tracking errors) of the controlled system. The third part is an adaptive control mechanism, which is to adapt some unknown constants of the least upper bounds of perturbations, so that the knowledge of the least upper bounds of matched and mismatched perturbations are not required. Several numerical examples and an application of controlling aircraft's velocity are demonstrated for showing the feasibility of the proposed control methodologies.

No description available.

Three robust tracking control design strategies are proposed in this dissertation for different classes of nonlinear MIMO dynamic systems with mismatched perturbations. The first controller design method is proposed for a class of nonlinear MIMO dynamic systems in canonical form. The second design procedure of controller is for the nonlinear MIMO dynamic systems without canonical form. A decentralized controller is presented in the last for perturbed large-scale systems with time-varying delay interconnections, where the knowledge of the exact function of time-delay is not required. These robust tracking controllers with a perturbation estimating scheme and an adaptive control mechanism embedded are designed by means of the variable structure control technique and Lyapunov stability theorem. The adaptive control mechanism is used to adapt the unknown upper-bound of perturbation estimation error, so that the knowledge of upper-bounds of perturbation as well as perturbation estimation error is not required. The chattering phenomenon is effectively alleviated, for the gain of the proposed controllers, which needs only to overcome the perturbation estimation error, is in general smaller than those of the traditional sliding mode controllers. Furthermore, the stability of the overall controlled systems is proved, and the desired tracking accuracy can be achieved by adjusting the design parameters of the proposed controller schemes. A numerical example for each controller's design is provided for demonstrating the feasibility of the proposed control schemes.

Abstract In this dissertation, four variable structure controllers are proposed for four different class of systems subjected to uncertainties and time varying delays respectively. In most cases, the variable structure control is incorporated with an adaptive law to drive the tracking error between the desired model and the controlled plant to zero. By using the Lyapunov stability theorem, the adaptive law is utilized for adapting the unknown upper bounds of the lumped perturbations so that the objective of asymptotical stability is achieved, and the variable structure control scheme is used for enhancing the robustness of stability of the controlled systems. Once the system enters the sliding region, the dynamics of controlled systems are insensitive to matching perturbations. It also shows that the proposed methodologies ensure the property of the globally uniformly ultimate boundness for the overall controlled system. Finally, four numerical examples are given for demonstrating the feasibility of the proposed control schemes.

Based on the Lyapunov stability theorem, an adaptive output feedback controller is proposed in this thesis for a class of multi-input multi-output (MIMO) dynamic systems with time-varying delay and disturbances. With an adaptive mechanism embeded in the proposed control scheme, the controller will automatically adapt the unknown upper bound of perturbation, so that the information of upper bounded of perturbations is not required. Once the controlled system reaches the switching hyperplane, not only the dynamics of system can be stabilized, but also the state trajectories can be driven into a small bounded region whose size can be adjusted through the design parameter. Two numerical examples are given for demonstrating the feasibility of the proposed control scheme.

Bioinformatics applications can address the transfer of information at several stages of the central dogma of molecular biology, including transcription and translation. This dissertation focuses on using Bayesian models to interpret biological data in bioinformatics, using Markov chain Monte Carlo (MCMC) for the inference method. First, we use our approach to interpret data at the transcription level. We propose a two-level hierarchical Bayesian model for variable selection on cDNA Microarray data. cDNA Microarray quantifies mRNA levels of a gene simultaneously so has thousands of genes in one sample. By observing the expression patterns of genes under various treatment conditions, important clues about gene function can be obtained. We consider a multivariate Bayesian regression model and assign priors that favor sparseness in terms of number of variables (genes) used. We introduce the use of different priors to promote different degrees of sparseness using a unified two-level hierarchical Bayesian model. Second, we apply our method to a problem related to the translation level. We develop hidden Markov models to model linker/non-linker sequence regions in a protein sequence. We use a linker index to exploit differences in amino acid composition between regions from sequence information alone. A goal of protein structure prediction is to take an amino acid sequence (represented as a sequence of letters) and predict its tertiary structure. The identification of linker regions in a protein sequence is valuable in predicting the three-dimensional structure. Because of the complexities of both models encountered in practice, we employ the Markov chain Monte Carlo method (MCMC), particularly Gibbs sampling (Gelfand and Smith, 1990) for the inference of the parameter estimation.

Empirical models of airflow output and power consumption were developed for series and parallel fan powered variable air volume terminal units at typical design pressure conditions. A testing procedure and experimental setup were developed to test sets of terminal units from three different manufacturers. Each set consisted of two series and two parallel units, each with 8 in. (203 mm) and 12 in. (304 mm) primary air inlets, for a total of four units in each set. Generalized models were developed for the series and parallel units, with coefficients varying by size and manufacturer. Statistical modeling utilized SAS software (2002). Fan power and airflow data were collected at downstream static pressures over a range from 0.1 to 0.5 in. w.g. (25 to 125 Pa) for the parallel terminal units. Downstream static pressure was held constant at 0.25 in. w.g. (62 Pa) for the series units. Upstream static pressures of all variable air volume (VAV) terminal units ranged from 0.1 to 2.0 in. w.g. (25 to 498 Pa). Data were collected at four different primary air damper positions. Data were also collected at four different terminal unit fan speeds, controlled by a silicon controlled rectifier (SCR). The models utilized the RMS voltage entering the terminal unit fan, the 'rake' sensor velocity pressure, and the downstream static pressure. In addition to the terminal unit airflow and power models, a model was developed to quantify air leakage in parallel terminal units, when the unit fan was off. In all but two of the VAV terminal units, the resulting models of airflow and power had R2 values greater than 0.90. In the two exceptions, there appeared to be manufacturing defects: either excessive air leakage or a faulty SCR that limited the effectiveness of the airflow and power models to capture the variation in the data.

In this dissertation three robust control strategies are proposed for a class of multi-input multi-output dynamic systems with matched or mismatched perturbations. Firstly, an adaptive variable structure observer and controller are introduced for solving the regulation problems, where some state variables are not measurable. By utilizing adaptive mechanisms in the design of sliding mode controller, one can enable the controlled systems not only to generate a reaching mode in finite time, but also to suppress the mismatched perturbations during the sliding mode. Secondly, the design of adaptive sliding mode controllers with application to robot manipulators is presented to solve the tracking problems. The dynamic equations of the controlled systems contain a perturbed leading coefficient matrix and can be either positive definite or negative definite. The asymptotical stability of the controlled systems will be attained if the proposed control scheme is employed. Thirdly, a design methodology of adaptive sliding mode controller based on T-S fuzzy model is proposed to solve tracking problems. It is shown that the trajectories of the controlled systems can be driven into a designated sliding surface in finite time, and the property of asymptotical stability is also guaranteed. All these three control schemes are designed by means of Lyapunov stability theorem. Each control scheme contains three parts. The first part is designed for eliminating measurable feedback signals. The second part is used for adjusting the convergent rate of state variables (or tracking errors) of the controlled system. The third part is the adaptive control mechanism, which is used to adapt some unknown constants of the least upper bounds of perturbations, so that the knowledge of the least upper bounds of matched or mismatched perturbations are not required. Several numerical examples and an application of controlling robot manipulator are demonstrated for showing the feasibility of the proposed control methodologies.

none

Floating-point multipliers are typically power hungry which is undesirable in many embedded applications. This paper proposes a variable-latency floating-point multiplier architecture, which is suitable for low-power, high-performance, and high-accuracy applications. The architecture splits the significand multiplier into upper and lower parts, and predicts the required significand product and sticky bit from upper part. In the case of correct prediction, the computation of lower part is disabled and the rounding operation is significantly simplified so that floating-point multiplication can be completed early. Finally, detailed design and simulation of the floating-point multiplier is presented, together with its evaluation by comparing power consumption with the fast and conventional floating-point multipliers. Experimental results demonstrate that the proposed double-precision multiplier consumes up to 26.41% and 24.97% less power and energy than the fast floating-point multiplier respectively at the expense of only small area and delay overhead. In addition, the results also show that the performance of proposed floating-point multiplier is very approximate to that of fast floating-point multipliers.