Biological Inspired Machine: Learning Based Mobile Robot Velocity Estimation

By

Acknowledgement

I would take this opportunity to thank my research supervisor, family and friends for their support and guidance without which this research would not have been possible.

DECLARATION

I, [type your full first names and surname here], declare that the contents of this dissertation/thesis represent my own unaided work, and that the dissertation/thesis has not previously been submitted for academic examination towards any qualification. Furthermore, it represents my own opinions and not necessarily those of the University.

Signed __________________ Date _________________

Abstract

In this study we try to explore the concept of “Biological Inspired Machine” in a holistic context. The main focus of the research is on “Biological Inspired Machine” and its relation with “learning based mobile robot velocity estimation”. The research also analyzes many aspects of “Biological Inspired Machine” and tries to gauge its effect on “learning based mobile robot velocity estimation”. Finally the research describes various factors which are responsible for “Biological Inspired Machine” and tries to describe the overall effect of “Biological Inspired Machine” on “learning based mobile robot velocity estimation”.

I. INTRODUCTION

THE PROBLEM of localization is central to endowing mobile machines with intelligence. Range sensors, such as sonar and ladar , , are particularly effective indoors due to many structural regularities such as flat walls and narrow corridors.

In the outdoors, these sensors become less robust given all the protrusions and surface irregularities . For example, a slight change in pose can result in large jumps in range reading because of tree trunks, moving branches, and leaves. Global Positioning System (GPS), coupled with other sensors or by itself , has also been extensively used. However, GPS may not be applicable in environments where there is no satellite visibility, such as under water, in caves, indoors, or on Mars. In those places, vision, which is our main perceptual system for localization, should be a viable alternative.

We first describe traditional vision localization techniques as background information to better demonstrate the advantages of using biological approaches. In Section I-B, we then introduce a robust biologically plausible vision system that concurrently observes a scene from two contrasting perspectives: Its rough overall layout (using gist) and detailed recognition only on select globally conspicuous locations (using saliency). In addition, Section I-C describes how using topological maps, which is analogous to how humans deal with spatial information, allows for a compact and accurate representation.

A. Traditional Vision-Based Localization

Existing vision-based localization systems can be categorized along several lines. The first one is according to image-view types, where some systems use ground-view images , and others use omnidirectional images , . Another categorization is according to localization goal, such as actual metric location or a coarser place or room number . Yet another grouping is according to whether or not the system is provided with a map or must build one as it locates itself.

The latter is known as Simultaneous Localization and Mapping (SLAM) One additional categorization to consider comes from the vision perspective, which classifies systems according to visual feature type, ...