Evolutionary Robotics: From Algorithms to Implementations

Citation

(2009), "Evolutionary Robotics: From Algorithms to Implementations", Industrial Robot, Vol. 36 No. 5. https://doi.org/10.1108/ir.2009.04936eae.004

Evolutionary Robotics: From Algorithms to Implementations

Article Type: Book review From: Industrial Robot: An International Journal, Volume 36, Issue 5

Lingfeng Wang, Kay Chen Tan and Chee Meng Chew,World Scientific,2006,$89,268 pp.,ISBN: 978-981-256-870-0,Web site: www.worldscibooks.com/compsci/6164.html

Creating autonomous robots is a primary goal of today's robotics. Achieving this goal is strongly related to the degree of intelligence embedded in the robot's controller. Evolutionary Robotics: From Algorthims to Implementations is one of the latest issues in robotics and evolutionary robotics (ER) is an emerging area of research for the automatic creation of autonomous robots. The book authored by Wang, Tan, and Chew constitutes a valuable reference to the concepts and methodologies related to ER. A collection of biological-inspired approaches for robot navigation, learning and control are presented with emphasis to genetic and evolutionary algorithms, artificial neural networks and fuzzy logic.

The book is organized in nine chapters. Each chapter ends with a summary and a broad bibliography with references for further study. Chapter 1 gives a comprehensive review on the variety biological-inspired approaches used in intelligent robot navigation, while future prospective and ideas in the field are suggested and discussed. Chapters 2 and 3 address the use of evolvable hardware in robotic systems as a means to adapt to changes in the environment. Chapter 2 surveys evolvable hardware in ER and presents characteristic experiments in the field of robot hardware evolution. Chapter 3 presents the design and real-time implementation of an evolvable hardware-based autonomous robot navigation system. Chapter 4 deals with sensor data fusion and learning for autonomous robot navigation in dynamic environments. In Chapter 5, a new task-oriented developmental learning approach for humanoid robots is presented. Using the notion of real-time experience, the proposed approach allows the robot to handle various tasks in unknown environments without the need of predefined the tasks.

The next three chapters are dedicated to the design and development of autonomous walking robots. In particular, Chapter 6 discusses a general robot control architecture based on a divide-and-conquer approach for bipedal walking. Chapter 7 presents a fuzzy logic controller tuned by a genetic algorithm for walking control of bipedal robots. Chapter 8 studies the application of genetic algorithms as optimization heuristics to optimize key parameters in the walking controlling of a humanoid robot. Chapter 9 concludes the book giving a synopsis of the most important developments presented in the previous eight chapters and discusses possible directions for future work in ER.

On the whole, the book is well structured and quite readable. All chapters are self-contained and readers can proceed directly with that they are interested. It is more suited to researchers and engineers working in the area of intelligent robotic systems. If more illustrative examples are added in each chapter, together with solved problems, practical questions and project exercises, it could be served as a text book in a graduate course on autonomous robots. Moreover, due to the large number of abbreviations used throughout the book, a table of abbreviations in an appendix will be quite useful for the reader.

In conclusion, this is an effective and interesting book that represents a valuable contribution to the field of biological-inspired algorithms and their applications in robotics. I definitely recommend the book to academic researchers, PhD students, engineers and professional practitioners interested in these topics.

Andreas C. NearchouFaculty of the Department of Business Administration, University of Patras, Patras, Greece