Multi-Agent Robotic Systems

Citation

(2002), "Multi-Agent Robotic Systems", Industrial Robot, Vol. 29 No. 6. https://doi.org/10.1108/ir.2002.04929fae.001

Multi-Agent Robotic Systems

Multi-Agent Robotic Systems

J. Liu and J. WuCRC Press2001304 pp.ISBN 0-8493-2288-X£62.99 (Hardback)

Keywords: Robotics

This book addresses the theoretical and experimental aspects of multi-agent robot systems. It will be of interest to those involved in machine intelligence and is suitable for students, engineers, researchers and practitioners in engineering and computer science.

“Multi-Agent Robotic System” contains 14 Chapters divided into five focal areas. Part I, Motivation, Approaches and Outstanding Issues, presents five chapters discussing the advantages and behavioural inspiration of multi-agent systems. Chapters 1 to 3 address Why Multiple Robots?; Toward Co-operative Control; and Approaches, respectively. Topics discussed include: co-operation-related research, design of multi-robot control, behaviour-based robotics, evolutionary robotics, and inspiration from biology and sociology. Reinforcement learning, genetic algorithms, artificial immune systems and probabilistic modelling, are amongst the topics presented in Chapter 4, Models and Techniques. Chapter 5, Outstanding Issues, discusses topics including: self organisation; local vs global performance; multi-robot learning; and emergent behaviour.

Part II, Case Studies in Learning, contains four chapters discussing the techniques, results, important aspects, and evolution of multi-agent reinforced learning. Topics presented include: autonomous group robots, group behaviours, collective sensing, initial spatial distribution, and evolving group motion strategies.

Part III, Case Studies in Adaptation, contains two chapters addressing Co-ordinated Manoeuvres in a Dual Agent System, and Collective Behaviour. Topics discussed include: dual agent learning, specialised roles in a dual-agent system, acquiring complex manoeuvres, collective box-pushing by applying repulsive forces, and convergence analysis for the fittest preserved evolution.

The final two parts of the book present Case Studies in Self-Organisation, and An Exploration Tool. Chapters 12 and 13 discuss Multi-Agent Self-Organisation, and Evolutionary Multi-Agent Self-Organisation, respectively, while Chapter 14 presents Toolboxes for Muli-Agent Robotics. The toolboxes described in this chapter can be downloaded from the book’s website (visit http://www.crcpress.com) and are a comprehensive collection of MATLAB functions written to assist research into multi- robot reinforced learning, adaptation, self- organisation, trajectory generation, and graphics.

“Multi-Agent Robotics Systems” is clearly illustrated and an excellent introduction into this fascinating area of robotics. It provides enough of the in-depth mathematical basis required by researchers in the field but is also descriptive enough to be of benefit to those who are relatively new to the field.