The relative merits of the Fisher linear discriminant function (Efron, 1975) and logistic regression procedure (Press and Wilson, 1978; McLachlan and Byth, 1979), applied to the two group discrimination problem under conditions of multivariate normality and common covariance, have been debated. In related research, DiPillo (1976, 1977, 1979) has argued that a biased Fisher linear discriminant function is preferable when one or more collinearities exist among the classifying variables. This paper proposes a generalized ridge logistic regression (GRL) estimator as a logistic analog to DiPillo's biased alternative estimator. Ridge and Principal Component logistic estimators proposed by Schaefer et al. (1984) for conventional logistic regression are shown to be special cases of this generalized ridge logistic estimator. Two Fisher estimators (Linear Discriminant Function (LDF) and Biased Linear Discriminant Function (BLDF)) and three logistic estimators (Linear Logistic Regression (LLR), Ridge Logistic Regression (RLR) and Principal Component Logistic Regression (PCLR)) are compared in a Monte Carlo simulation under varying conditions of distance between populations, training set s1ze and degree of collinearity. A new approach to the selection of the ridge parameter in the BLDF method is proposed and evaluated. The results of the simulation indicate that two of the biased estimators (BLDF, RLR) produce smaller MSE values and are more stable estimators (smaller standard deviations) than their unbiased counterparts. But the improved performance for MSE does not translate into equivalent improvement in error rates. The expected actual error rates are only marginally smaller for the biased estimators. The results suggest that small training set size, rather than strong collinearity, may produce the greatest classification advantage for the biased estimators. The unbiased estimators (LDF, LLR) produce smaller average apparent error rates. The relative advantage of the Fisher estimators over the logistic estimators is maintained. But, given that the comparison is made under conditions most favorable to the Fisher estimators, the absolute advantage of the Fisher estimators is small. The new ridge parameter selection method for the BLDF estimator performs as well as, but no better than, the method used by DiPillo. The PCLR estimator shows performance comparable to the other estimators when there is a high level of collinearity. However, the estimator gives up a significant degree of performance in conditions where collinearity is not a problem. / Graduation date: 1991

The internship company does not have a standard procedure for preparing an engineer’s estimate of probable construction cost document (engineer’s estimate) for municipal projects. Every project manager employs a methodology that is a slightly different variation of the historical data approach. The internship objective was to develop a construction unit price estimation model that provides more accurate results than the company’s existing unit price estimation methodology for the City of Fort Worth construction projects. To accomplish the internship objective several tasks were conducted, including; gathering City of Fort Worth construction projects bid tabulation data (including all bids) for the past three years; developing three construction item unit price databases using the data collected; conducting statistical analyses using the unit price databases;developing tables and graphs showing the construction cost items and their appropriate estimated unit prices to be used by the project managers in their cost estimates; developing an approach to apply construction unit costs which adjusts for unique project characteristics; developing guidelines for using the developed tables and graphs to estimate unit prices for municipal projects; using one recent project to compare the company’s existing unit price estimation methodology and the new developed model with actual unit bid prices; and developing guidelines for updating the unit price database, tables, and graphs. The study made use of both normal and log-normal distributions to model the unit bid price data collected from the City of Fort Worth. The factors that are perceived to influence a contractor’s unit bid price for a given item were identified and given a degree of impact on the project by the project managers. The factor that had the highest impact on the unit bid prices was discovered to be item quantity. The unit price estimating methodology presented in this study generated a better fit than the internship company’s original method for predicting the actual average unit bid prices for the one case study the methodology was applied.

In this thesis, we try to provide a broad econometric analysis of a class of risk measures, distortion risk measures (DRM). With carefully selected functional form, the Value-at-Risk (VaR) and Tail-VaR (TVaR) are special cases of DRMs. Besides, the DRM also admits interpretation in the sense of non-expected utility type of preferences. We first provide a unified statistical framework for the nonparametric estimators of the DRMs in a univariate case. The asymptotic properties of both the DRMs and their sensitivities with respect to the parameters representing risk aversion and/or pessimism are derived. Moreover, the relationships between the VaR and TVaR are also investigated in detail, which, we hope, can shed new lights on the way passing one risk measure to another. Then, the analysis of DRMs are extended to a multi-dimensional framework, where the DRM is computed for a portfolio consisting of many primitive assets. Analogous to the mean-variance frontier analysis, we study the efficient portfolio frontier when both objective and constraint are replaced by the DRMs. We call this the DRM-DRM framework. Under a nonparametric setting, we propose three asymptotic test statistics for evaluating the efficiency of a given portfolio. Finally, we discuss the criteria used for evaluating models used to forecast the VaRs. More precisely, we propose a criterion which takes into account the loss levels beyond the VaRs.

Today, a rapid development towards fuel efficient technological aids for vehicles is in progress. One step towards this is the development of platooning systems. The main concept of platooning is to let several heavy duty vehicles (HDVs) drive in a convoy and share important information with each other via wireless communication. This thesis describes one out of three subsystems in a project developed to handle the process from raw sensor data to control signal. The goal of the project is to achieve a safe and smooth control with the main purpose of reduced fuel consumption. This subsystem processes the raw sensor data received from the different HDVs. The purpose is to estimate the positions and velocities of the vehicles in a platoon, taking into account that packet-loss, out of sequence measurements and irrelevant information can occur. This is achieved by filtering the information from different sensors in an Extended Kalman Filter and converting it into a local coordinate system with the origin in the ego vehicle. Moreover, the estimates are sorted and categorized into classes with respect to the status of the vehicles. The result of the thesis is useful estimates that are independent of outer effects in a local reference system with origin in the host vehicle. This information can then be used for further sensor fusion and implementation of a Model Predictive Controller (MPC) in two other subsystems. These three subsystems result in a smooth and safe control with an average reduced fuel consumption of approxi- mately 11.1% when the vehicles drive with a distance of 0.5 seconds in a simulated environment. / Dagens utveckling inom fordonsindustrin fokuserar mer och mer påutveckling av bränsleeffektiva hjälpmedel. Ett steg i denna riktning är utvecklingen av platooningsystem. Huvudkonceptet med platooning är att låta flera tunga fordon köra i följd i en konvoj och dela viktig information med varandra via trådlös kommuni- kation och en automatiserad styrstrategi. Detta examensarbete beskriver ett utav tre delsystem i ett projekt som är utvecklat för att hantera en process från rå sensordata till styrsignaler för fordonen. Målet är att uppnå en säker och mjuk reglering med huvudsyftet att reducera bränsleförbrukningen. Det här delsystemet behandlar mottagen sensordata från de olika fordonen. Målet med delsystemet är att skatta positioner och hastigheter för fordonen i konvojen med hänsyn till att förlorad, försenad eller irrelevant information från det trådlösa nätverket kan förekomma. Detta uppnås genom filtrering i ett Extended Kalman Filter och konvertering till ett lokalt referenssystem med origo i det egna fordo- net. Utöver detta sorteras informationen och kategoriseras in i olika klasser efter fordonens status. Examensarbetet resulterade i användbara skattningar oberoende av yttre om- ständigheter i ett lokalt referenssystem med origo i det egna fordonet. Denna information kan användas vidare för ytterligare sensorfusion och implementering av en modellbaserad prediktionsregulator (MPC) i två andra delsystem. De tre delsystemen resulterade i en mjuk och säker reglering och en reducerad bränsleför- brukning med i genomsnitt 11.1% då fordonen körde med 0.5 sekunders avstånd i en simulerad miljö.

This thesis proposes a new hybrid particle swarm optimizer, which employs landscape estimation and the cooperative behavior of different particles to significantly improve the performance of the original algorithm. The landscape estimation is to explore the landscape of the function in order to predict whether the function is unimodal or multimodal. Then we can decide how to optimize the function accordingly. The cooperative behavior is achieved by using two swarms, in which one swarm explores only a single dimension at a time, and the other explores all dimensions simultaneously. Furthermore, we also propose a movement tracking-based strategy to adjust the maximal velocity of the particles. This strategy can control the exploration and exploitation abilities of the swarm efficiency. Finally, we testify the performance of the proposed approach on a suite of unimodal/multimodal benchmark functions and provide comparisons with other recent variants of the PSO. The results show that our approach outperforms other methods in most of the benchmark problems.

Battery is the sole electrical energy source when electric vehicle(EV) is moving. To reduce traveling anxiety, an effective energy management system to indicate the state-of-charge (SOC) of the battery and make a balance between vehicle performance and endurance is very important. This research is aimed to develop a SOC estimation system with high accuracy. The proposed method in this thesis is based on under load voltage and multilevel Peukert's equation to estimate the SOC. The proposed method is compared with the open circuit voltage method for initial SOC estimation and with coulometric method for cumulative SOC estimation under various EV driving conditions simulated by an adjustable electronics load. Experimental results indicate that the proposed method can provide reasonable accuracy as compared with other tested methods for LiFePO4 battery SOC estimations.

Flow speed describes general traffic operation conditions on a segment of roadway. It is also used to diagnose special conditions such as congestion and incidents. Accurate speed estimation plays a critical role in traffic management or traveler information systems. Data from loop detectors have been primary sources for traffic information, and single loop are the predominant loop detector type in many places. However, single loop detectors do not produce speed output. Therefore, speed estimation using single loop outputs has been an important issue for decades. This dissertation research presents two methodologies for speed estimation using single loop outputs. Based on findings from past studies and examinations in this research, it is verified that speed estimation is a nonlinear system under various traffic conditions. Thus, a methodology of using Unscented Kalman Filter (UKF) is first proposed for such a system. The UKF is a parametric filtering technique that is suitable for nonlinear problems. Through an Unscented Transformation (UT), the UKF is able to capture the posterior mean and covariance of a Gaussian random variable accurately for a nonlinear system without linearization. This research further shows that speed estimation is a nonlinear non-Gaussian system. However, Kalman filters including the UKF are established based on the Gaussian assumption. Thus, another nonlinear filtering technique for non-Gaussian systems, the Particle Filter (PF), is introduced. By combining the strengths of both the PF and the UKF, the second speed estimation methodology—Unscented Particle Filter (UPF) is proposed for speed estimation. The use of the UPF avoids the limitations of the UKF and the PF. Detector data are collected from multiple freeway locations and the microscopic traffic simulation program CORSIM. The developed methods are applied to the collected data for speed estimation. The results show that both proposed methods have high accuracies of speed estimation. Between the UKF and the UPF, the UPF has better performance but has higher computation cost. The improvement of speed estimation will benefit real-time traffic operations by improving the performance of applications such as travel time estimation using a series of single loops in the network, incident detection, and large truck volume estimation. Therefore, the work enables traffic analysts to use single loop outputs in a more cost-effective way.

Conventional Cobb-Douglas and Transcendental Logarithmic production functions widely used in Stochastic Production Frontier Estimation and Inefficiency Analysis have merits and deficiencies. The Cobb-Douglas function imposes monotonicity and concavity constraints required by microeconomic theory. However it is inflexible and implies undesired assumptions as well. The Trans-log function is very flexible and does not imply undesired assumptions, yet it is very hard to impose both monotonicity and concavity constraints. The first essay introduced a class of stochastic production frontier estimation models that impose monotonicity and concavity constraints and suggested models that are very flexible. Researchers can use arbitrary order of polynomial functions or any function of independent variables within the suggested frameworks. Also shown was that adopting suggested models could greatly increase predictive accuracy through simulations. In the second essay we generalized the suggested models with the Inefficiency Analysis technique. In the last essay we extended the models developed in the previous two essays with regression spline and let the data decide how flexible or complicated the model should be. We showed the improvement of deterministic frontier estimation this extension could bring through simulations, as well. Works in this dissertation reduced the gap between conventional structural models and nonparametric models in stochastic frontier estimation field. This dissertation offered applied researchers Stochastic Production Frontier models that are more accurate and flexible than previous ones. It also preserves constraints of economic theory.

In this thesis, a precoding scheme is proposed for channel estimation in orthogonal frequency-division multiplexing (OFDM) systems. The precoding scheme utilizes a special code matrix before the inverse fast Fourier transform (IFFT) at the transmitter. The row vectors of the matrix have constant amplitudes in both time domain and frequency domain. With the prcoding scheme, a semi-blind channel estimation method is proposed by using the characteristics of the row code sequences. In the proposed scheme, the channel frequency responses of all sub-carriers can be obtained by using only one pilot sub-carrier, and the proposed architecture can not only increase the data rate, but also avoid interpolation error in channel estimation. In addition, the normalized mean square error (NMSE) function is derived and parameters are optimized to improve system performance. The proposed precoding architecture and channel estimation scheme are shown to have better performances in bit error rate by conducting computer simulation experiments.

In this thesis, a modified Kalman filter is proposed for time varying channel estimation in orthogonal frequency division multiplexing (OFDM). The proposed scheme adopts pre-coding scheme and minimum mean squared error (MMSE) equalizer to improve system performance. By using pre-coding schemes, information can be protected by signal diversities, which prevent Kalman filter to disperse due to erroneous data signals. In this investigation, the proposed system architecture is verified by using simulation experiments. Simulation results demonstrate that the proposed schemes substantially improve system performances under verious channel conditions.