Pandya, Abhijit S.

A neural network based model for prediction of bridge condition rating is proposed. The back-propagation algorithm is used to train the network to recognize the pattern of deterioration of bridges and use this knowledge in predicting the future condition rating of a bridge. The various factors which influence the deterioration rate are considered as input to the system. The model then predicts the condition rating of the three major sub-components of a bridge viz. the deck, sub-structure and the super-structure. Fuzzy logic is used to evaluate the overall condition rating of the bridge using the condition rating of the components. To demonstrate the superiority of the neural network model over the traditional models, the history of the deterioration rates for the components were also considered in the prediction of their future condition. The proposed system is versatile and can be easily extended to include other parameters and updated from time to time without much effort.

The last decade has seen a surge in demand for cellular and WLAN networks. With the introduction of Voice Over IP, cellular companies are looking at WLAN-Cellular integrated networks that shall provide users with economical voice and data services. These networks shall be complimentary to the existing cellular networks. A lot of money is spent in registering and authenticating new users, since they are separately authenticated and registered on the WLAN and Cellular domains. This leads to extra costs for the company. Thus for the integrated networks to have an impact on the market some issues such as simpler authentication and registration must be resolved. Therefore we propose a new inter-working model that shall addresses the authentication and registration problem for an integrated network for voice and data. The Single authentication system of the new inter-working model, shall authenticate the user in an integrated network using the SIM credentials, this authentication shall be valid for both voice and data. Also registration costs will be saved by preventing separate registration of users in the WLAN and Cellular domain.

There is a mushrooming demand for battery operated applications that require intensive computation in portable environments. This has motivated the research and development of techniques that reduce power in CMOS digital circuits while maintaining their computational throughput. The two essentials to achieve a low power design are miniaturization and long battery life. Lowering the supply voltage is one of the most effective ways to achieve low-power performance as power dissipation in digital CMOS circuits is approximately proportional to the square of supply voltage. The basic idea behind this thesis is that it proposes new designs of transfer gate based logical circuits, which use lower supply voltage and less number of transistors than the conventional designs. This work evaluates the obtained results from the proposed designs of the low-power ALU with that from the standard CMOS, other low power designs namely, Wang's XOR, XNOR and Inverter based gates. It was observed that the proposed designs perform better in terms of power consumption than the standard CMOS designs, and the other low power designs mentioned above.

In this thesis we present an intelligent forecaster based on neural network technology to capture the future path of the market indicator. This thesis is about the development of a new methodology in financial forecasting. An effort is made to develop a neural network forecaster using the financial indicators as the input variables. A complex recurrent neural network is used to capture the behavior of the nonlinear characteristics of the S&P 500. The main outcome of this research is, a systematic way of constructing a forecaster for nonlinear and non-stationary data series of S&P 500 that leads to very good out-of-sample prediction. The results of the training and testing of the network are presented along with conclusion. The tool used for the validation of this research is "Brainmaker". This thesis also contains a brief survey of available tools for financial forecasting.

The goal of time series forecasting is to identify the underlying pattern and use these patterns to predict the future path of the series. To capture the future path of a dynamic stock market variable is one of the toughest challenges. This thesis is about the development of a new methodology in financial forecasting. An effort is made to develop a neural network forecaster using time-series phenomena. The main outcome of this new approach for financial forecasting is a systematic way of constructing a Neural Network Forecaster for nonlinear and non-stationary time-series data that leads to very good out-of-sample prediction. The tool used for the validation of this research is "Brainmaker". This thesis also contains a small survey of available tools used for financial forecasting.

Software reuse has been looked upon in recent years as a promising mechanism for achieving increased levels of software quality and productivity within an organization. A form of software reuse which has been gaining in popularity is the use of design patterns. Design patterns are a higher level of abstraction than source code and are proving to be a valuable resource for both software developers and new hires within a company. This thesis develops the idea of applying design patterns to the Computer Aided Design (CAD) software development environment. The benefits and costs associated with implementing a software reuse strategy are explained and the reasoning for developing design patterns is given. Design patterns are then described in detail and a potential method for applying design patterns within the CAD environment is demonstrated through the development of a CAD design pattern catalog.

This project illustrates the neural network approach to constructing a fuzzy logic decision system. This technique employs an artificial neural network (ANN) to recognize the relationships that exit between the various inputs and outputs. An ANN is constructed based on the variables present in the application. The network is trained and tested. Various training methods are explored, some of which include auxiliary input and output columns. After successful testing, the ANN is exposed to new data and the results are grouped into fuzzy membership sets based membership evaluation rules. This data grouping forms the basis of a new ANN. The network is now trained and tested with the fuzzy membership data. New data is presented to the trained network and the results form the fuzzy implications. This approach is used to compute skid resistance values from G-analyst accelerometer readings on open grid bridge decks.

Accurately predicting the quality of software is a major problem in any software development project. Software engineers develop models that provide early estimates of quality metrics which allow them to take action against emerging quality problems. Most often the predictive models are based upon multiple regression analysis which become unstable when certain data assumptions are not met. Since neural networks require no data assumptions, they are more appropriate for predicting software quality. This study proposes an improved neural network architecture that significantly outperforms multiple regression and other neural network attempts at modeling software quality. This is demonstrated by applying this approach to several large commercial software systems. After developing neural network models, we develop regression models on the same data. We find that the neural network models surpass the regression models in terms of predictive quality on the data sets considered.

A barrier to the use of digital imaging is the vast storage requirements involved. One solution is compression. Since imagery is ultimately subject to human visual perception, it is worthwhile to design and implement an algorithm which performs compression as a function of perception. The underlying premise of the thesis is that if the algorithm closely matches visual perception thresholds, then its coded images contain only the components necessary to recreate the perception of the visual stimulus. Psychophysical test results are used to map the thresholds of visual perception, and develop an algorithm that codes only the image content exceeding those thresholds. The image coding algorithm is simulated in software to demonstrate compression of a single frame image. The simulation results are provided. The algorithm is also adapted to real-time video compression for implementation in hardware.

The design of an Autonomous Undersea Vehicle (AUV) control system is a significant challenge in-light of the highly uncertain nature of the ocean environment together with partially known nonlinear vehicle dynamics. This thesis describes a Neural Network architecture called Cerebellar Model Arithmetic Computer (CMAC). CMAC is used to control a model of an Autonomous Underwater Vehicle. The AUV model consists of two input parameters, the rudder and stern plane deflections, controlling six output parameters; forward velocity, vertical velocity, pitch angle, side velocity, roll angle, and yaw angle. Properties of CMAC and results of computer simulations for identification and control of the AUV model are presented.