Year

2011

Degree Name

Master of Computer Science - Research

Department

School of Computer Science and Software Engineering

Abstract

Recognising objects using sonar presents a significant research problem, especially when echo information for feature extraction is scarce. Blind people, however, have shown an amazing ability to identify such objects using an ultrasonic mobility aid. For example a blind man named Fred Gissoni was able to recognise a table and chair in a restaurant in order to find a place to eat. Bats have also demonstrated an amazing capacity to navigate in dark caves using ultrasonic echolocation. Given these remarkable feats, there must be more information in the signal of an echo than previously thought. If these features could be identified a more robust approach to recognising objects using sonar could be developed. This research attempts to address this issue using a framework based on Mahalanobis distance. As a particular example the echoes of a table and chair are studied.

FoR codes (2008)

0801 ARTIFICIAL INTELLIGENCE AND IMAGE PROCESSING, 080101 Adaptive Agents and Intelligent Robotics, 080109 Pattern Recognition and Data Mining, 090609 Signal Processing

Download

Share

COinS
 

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.