Multiple Abstraction Hierarchies For Mobile Robot Operation In Large Environments

by Galindo, Cipriano; Fernandez-madrigal, Juan-antonio; Gonzalez, Javier

• ISBN13: 9783540726883
• ISBN10: 3540726888
• Format: Hardcover
• Copyright: 2007-09-01
• Publisher: Springer Verlag

Summary

This book focuses on the efficient performance of mobile robots through the use of multi-hierarchical symbolic representations of the environment.A mobile robot intended to perform deliberative actions must possess some symbolic representation of its workspace, but such representations of real environments usually become so large that should be conveniently arranged in order to facilitate and, in some cases, make possible their use.Apart from the drawback of dealing with large amounts of information, other problems stand out when using symbolic representations. One is keeping the model properly optimized with respect to the robot tasks while maintaining it coherent with reality. Another one is the creation (or modification) of symbols from sensorial data.This book addresses all these issues considering symbolic multi-hierarchical structures. Such structures, based on the concept of abstraction, allow a robot to speed up its operation in large environments. Practical solutions tested on real robots (for instance, a robotic wheelchair for elder people) are provided.The book is intended for PhD students and in general for robotics, computer science, and artificial intelligence researchers.

Table of Contents

A Robotic Future	p. 1
Why is it Not Real?	p. 2
Making it Possible	p. 5
Outline	p. 7
Multi-Hierarchical, Symbolic Representation of the Environment	p. 9
Introduction	p. 10
Informal Description of a Multi-Hierarchical Model Based on Graphs	p. 13
What is an AH-Graph?	p. 13
What is a Multi-AH-Graph?	p. 14
Formalization of Graphs	p. 15
Formalization of Graph Abstraction	p. 18
Category Theory for Abstraction and Refinement of Graphs	p. 20
The Category of Graphs with Abstraction	p. 20
The Category of Graphs with Refinements	p. 25
Conclusions	p. 27
Application of Hierarchical Models to Mobile Robot Task Planning	p. 29
Introduction	p. 30
Formalization of Classical Planning in Category Theory	p. 33
Planning Basics	p. 34
The Category of Planning States with Abstraction	p. 35
Functors Between the AGraph and the AState Categories	p. 37
Hierarchical Planning with CV AGraph and AState	p. 42
Hierarchical Planning Through Plan Guidance	p. 47
Hierarchical Planning Through Action Guidance	p. 54
Anomalies in Hierarchical Planning	p. 57
Experimental Results	p. 58
Conclusions	p. 61
Mobile Robot Operation with Multi-Hierarchies	p. 65
Introduction	p. 65
A Multi-Hierarchical World Model for a Mobile Robot	p. 68
Task-Planning Hierarchies	p. 68
Localization Hierarchy	p. 69
Cognitive Hierarchy	p. 71
The Utility of the Multi-Hierarchical Model for Human-Robot Interaction	p. 73
The Inter-Hierarchy Translation Process	p. 75
Interactive Task-Planning	p. 77
Conclusions	p. 83
Automatic Learning of Hierarchical Models	p. 85
Introduction	p. 86
The ELVIRA Framework	p. 88
Model Creation/Maintenance	p. 89
Human-Assisted Symbolic Modeling	p. 90
Perceptual Anchoring	p. 91
Model Optimization	p. 95
Evolutionary Hierarchy Optimizer	p. 96
Population	p. 97
Individual Recombination	p. 99
Individual Mutation	p. 101
Individual Evaluation	p. 104
Memory and Time Consumption	p. 105
Experimental Results	p. 107
Real Evaluation	p. 107
Simulated Experiments	p. 111
Considerations for Outdoor or Dynamic Environments	p. 113
Conclusions	p. 114
Implementation and Experiences on a Real Robot	p. 115
Introduction	p. 116
Related Works on Human-Robot Interaction	p. 119
Overview of the ACRHIN Architecture	p. 120
The Deliberative Level	p. 123
Symbolic Management	p. 123
Task Manager	p. 126
The Executive Level	p. 126
Alert System	p. 127
Plan Executor and Alert Manager (PLEXAM)	p. 129
The Functional Level	p. 129
Experiences on the SENA Robotic Wheelchair	p. 131
The SENA Robotic Wheelchair	p. 131
Software Details	p. 134
Experiences on SENA	p. 135
Conclusions	p. 141
Conclusions	p. 143
Appendices	p. 145
Mathematical Demonstrations for the Formalization of the Graph Category	p. 145
Composition of Abstractions in AGraph is an Abstraction	p. 145
Composition of Abstractions in AGraph is Associative	p. 148
Composition and Identity of Abstractions in AGraph	p. 149
Composition of Refinements in RGraph is a Refinement	p. 150
Composition of Refinements in RGraph is Associative	p. 153
Composition and Identity of Refinements in RGraph	p. 154
Demonstration that [Psi] is a Functor from CAVGraph to AState	p. 157
[Psi] Preserves Identity	p. 157
[Psi] Preserves Composition of Arrows	p. 158
Planning Domain	p. 161
References	p. 167
Index	p. 177
Table of Contents provided by Ingram. All Rights Reserved.