CEC Theses and Dissertations

Title

Intelligent Collision Warning System Based on Fuzzy Logic and Neural Network Technologies

Date of Award

1997

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Graduate School of Computer and Information Sciences

Advisor

Wilker Shane Bruce

Committee Member

Junping Sun

Committee Member

S. Rollins Guild

Abstract

The recent technological changes in computer and industrial control systems have been steadily extending the capabilities to handle a broad range of complex systems. The emergence and development of computer technology and intelligent systems during the past few decades have created a highly promising direction in the field of artificial intelligence. It is increasingly difficult to describe any real system as the level of complexity continues to increase. A combination of systems and techniques are necessary to solve many complex problems. This new direction involves the use of fuzzy logic and artificial neural network theory to enhance the ability of intelligent systems that can learn from experience and to adapt to changes in an environment of uncertainty and imprecision.

The Intelligent Automotive Collision Warning System was developed as a rule based system by integrating a fuzzy logic controller with artificial neural network software. The Intelligent Automotive Collision Warning system constantly monitors the speed of the vehicle and the distance of any object in front of the vehicle using an ultrasonic ranging module to warn the operator to maintain a safe operating distance by using fuzzy logic theory and artificial neural network software.

Descriptive statistics was used for collecting and organizing the data. Inferential statistics was used to prove the hypotheses based on the results of the collected data. NeuFuz4 software was used to train the neural network and to optimize the fuzzy rule base. The fuzzy logic technology provided a means of converting a linguistic control strategy to operate the warning system. The input/output relationship was defined by fuzzy membership functions which enabled the numerical inputs to be expressed as fuzzy variables using linguistic terms. A new fuzzy logic operator was also developed to optimize the fuzzy input/output relationship.

This document is currently not available here.

  Link to NovaCat

Share

COinS