Space-Time Continuous Models of Swarm Robotic Systems (eBook)

Supporting Global-to-Local Programming

Heiko Hamann (Autor)

eBook Download: PDF

2010 | 2010
160 Seiten
Springer Berlin (Verlag)
978-3-642-13377-0 (ISBN)

Infos

In this book, a generic model in as far as possible mathematical closed-formis developed that predicts the behavior of large self-organizing robot groups (robot swarms) based on their control algorithm. In addition, an extensive subsumption of the relatively young and distinctive interdisciplinary research field of swarm robotics is emphasized. The connection to many related fields is highlighted and the concepts and methods borrowed from these fields are described shortly.

Inhalt

Title Page 2
Abstract 5
Contents 7
Introduction and Purpose 11
Objectives 12
Approach 13
Outline 13
Fundamentals of Swarm Robotics – An Interdisciplinary Approach 14
Definition of Fundamental Concepts 14
Agent 14
Agent–Agent Interaction 15
Macroscopic and Microscopic Level 15
Phenomenological Approach 15
Self-organization 16
Emergence 16
Stigmergy 18
Micro-Macro Link 18
Swarm Intelligence 18
From Natural to Artificial Swarms 19
Swarms as Problem Solvers 19
Robotics 20
Mobile Robotics 20
Micro-Robotics 20
Swarm Robotics 21
Embodied Cognitive Science 24
Sensor/Actuator Networks 24
Software Concepts for Distributed Systems 25
(Distributed) Artificial Intelligence and Multi-robot Systems 25
Multi-Agent Systems 25
Artificial Life 26
Amorphous Computing 26
Self-awareness in Distributed Systems 27
Science of Self-organization 27
Theory of Dissipative Structures 27
Synergetics 28
Philosophical Issues 29
Emergence and Novelty 29
Multi-particle Systems and Computation 33
Summary 34
State-of-the-Art in Modeling and Design of Swarms 36
Modeling Swarm Behavior and Collective Behavior 36
Agent-Based Modeling 36
Classical Approach – Control Theory 37
Swarm Robotics 37
Physics 38
Chemistry 39
Biology 41
Micro, Macro, and Stochastic Modeling – Road Traffic 42
Design of Emergent Behavior 44
Global-to-Local Programming – Engineering of Emergence 44
Antagonism of Concepts 44
Programming “by Hand” 45
Automatic and Semi-automatic Methods 47
Summary 48
A Framework of Models for Swarm Robotic Systems 49
Modeling Multi-particle Systems – Example of Brownian Motion 49
Introduction – History of Brownian Motion 50
A Microscopic Model – The Langevin Equation 50
A Macroscopic Model – The Fokker–Planck Equation 55
Discussion of the Methodology 64
Purpose of a Model in the Design of Swarm Algorithms 65
Benefits of a Symbolic Model 66
Consideration of Alternative Methods 67
A Physically Motivated Approach 68
Modeling Robot Motion 69
Modeling Robot-Robot Interactions 73
External Influence – Modeling the Environment 73
Challenge of Modeling Communication 74
Towards General Methodological Principles 75
Relevance to the Concept of Computation 76
From Swarm Behavior to World-Embedded Computation 76
The Concept of Embodied Computation 77
Discussion 79
Discussion of Limitations 79
Discussion of Benefits 81
Discussion of the Relevance to Computation 82
Summary 83
Validation by Results of Experiments and Simulations 85
Collision-Based Adaptive Swarm Aggregation 85
The Swarm Robot “Jasmine” 86
Scenario 87
Model 89
Results 92
Discussion 95
Collective Perception 97
Scenario 97
Model 100
Results and Discussion 104
Emergent Taxis 108
Scenario 108
Model 109
Results 112
Foraging 113
Scenario 113
Model 115
Results 117
Robot Aggregation as a Form of Computation 121
The Euclidean Steiner Tree Problem 122
Growing Random Trees 123
Results 126
Discussion 129
Summary 129
Conclusion and Outlook 131
Conclusion 131
Outlook 132
Acknowledgments 133
Appendix A Numerical Simulation of the Langevin Equation 134
Appendix B Simple Numerical Solver of the Fokker–PlanckEquation 136
References 138
Index 150

Textauszug

"Chapter 5 Validation by Results of Experiments and Simulations (p. 77-78)

Several case studies were carried out to approve the utility and to validate the accurateness of the proposed model. One of these studies is based on experiments with real robots the other studies were performed using simulators. Each of the following examples is a little research project of its own although the model framework is a great support. There is still a respectable amount of creativity and insight necessary to obtain a satisfying result. The aim of these case studies is not to compare the swarm algorithms to alternatives but rather to take them for granted and to check what can be predicted and with what accuracy.

5.1 Collision-Based Adaptive Swarm Aggregation

The swarm algorithm, that is investigated in this case study, was designed and experimentally evaluated by Schmickl et al. (2008). For the experiments 15 Jasmine swarm robots were used, which have been developed at the Institute for Process Control and Robotics here at the Universit¨at Karlsruhe (TH) in cooperation with the University of Stuttgart (Kornienko et al., 2005a; Szymanski and W¨orn, 2007).

In this scenario, these robots were controlled by a bio-inspired algorithm, which is derived from honeybees’ navigation behavior in a temperature gradient. The following speci?c modeling approach and the results were previously reported in a short form in Hamann and W¨orn (2008b), Hamann et al. (2008), and Schmickl et al. (2009). Further studies with focus on the symmetry breaking capabilities of this system are reported in Hamann et al. (2010).

5.1.1 The Swarm Robot “Jasmine”

The swarm robot Jasmine (see Figure 5.1 and Jasmine (2008)) was developed especially for swarm robot research. Despite its small size of about 30×30×30 mm3, it has good local communication abilities and a far distance scanning and distance measuring sensor. The good communication abilities result from six infrared sensors and emitters arranged around the robot with a displacement of 60 degrees. These sensors are also used for short distance measurements.

The far distance measuring sensor is hooked to the front of the robot. Two differentially driven wheels give this robot a high manoeuvrability at a high speed. Generation three of the robot, that was used here, has an Atmel Mega 168 micro-controller with 1 Kbyte RAM and 16 Kbyte Flash. A single LiPo battery pack is suf?cient for up to two hours of motion and optical encoders are available that allow odometric measurements in the mm-range."

Produkt-Details

• Erscheint lt. Verlag: 18.5.2010
• Reihe/Serie: Cognitive Systems Monographs
• Reihe/Serie: Cognitive Systems Monographs
• Zusatzinfo: 160 p. 42 illus. in color.
• Verlagsort: Berlin
• Sprache: englisch
• Themenwelt: Geisteswissenschaften
• Themenwelt: Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik
• Themenwelt: Technik ► Elektrotechnik / Energietechnik
• Themenwelt: Technik ► Maschinenbau
• Schlagworte: algorithms • Cognitve Systems • Complexity • Control Algorithm • Modeling • programming • robot • Robotics • Self-Organization • Simulation • swarm robotics
• ISBN-10: 3-642-13377-0 / 3642133770
• ISBN-13: 978-3-642-13377-0 / 9783642133770