Evolutionary Robotics: Fundamentals and Applications

By Fouad Sabry

What Is Evolutionary Robotics

An embodied approach to artificial intelligence (AI), evolutionary robotics involves the creation of robots by the use of Darwinian concepts of natural selection during the design process. A behavioral goal can be used to guide the design of a robot or a subsystem of a robot, such as a neural controller. This design is then optimized. In most cases, designs are reviewed using computer simulations because it would be extremely expensive to physically fabricate thousands or millions of designs and test them in the actual world due to the amount of time, money, and risk involved.

How You Will Benefit

(I) Insights, and validations about the following topics:

Chapter 1: Evolutionary Robotics

Chapter 2: Evolutionary Algorithm

Chapter 3: Evolutionary Computation

Chapter 4: Neuroevolution

Chapter 5: Neuroevolution of Augmenting Topologies

Chapter 6: Evolvable Hardware

Chapter 7: Dario Floreano

Chapter 8: Stefano Nolfi

Chapter 9: HyperNEAT

Chapter 10: Evolutionary Developmental Robotics

(II) Answering the public top questions about evolutionary robotics.

(III) Real world examples for the usage of evolutionary robotics in many fields.

(IV) 17 appendices to explain, briefly, 266 emerging technologies in each industry to have 360-degree full understanding of evolutionary robotics' technologies.

Who This Book Is For

Professionals, undergraduate and graduate students, enthusiasts, hobbyists, and those who want to go beyond basic knowledge or information for any kind of evolutionary robotics.

• Language: English
• Publisher: One Billion Knowledgeable
• Release date: Jul 7, 2023

Book preview

Chapter 1: Evolutionary robotics

An embodied approach to artificial intelligence (AI), evolutionary robotics involves the creation of robots by the use of Darwinian concepts of natural selection throughout the design process. The construction of a robot or a component of a robot, such as its neural controller, is optimized in relation to the achievement of a certain behavioral objective (e.g. run as fast as possible). Simulations are often used to assess ideas since physically constructing dozens or millions of designs and testing them in the actual world is impractically costly in terms of both time and money, not to mention the risk involved.

An experiment in evolutionary robotics often begins with a population of robot designs that have been produced at random. The designs that perform the poorest are thrown out, and their places are taken by mutations and/or combinations of the designs that do better. This evolutionary algorithm will keep running until either a certain length of time has elapsed or some performance measure that was set as a goal has been exceeded.

Methods from the field of evolutionary robotics are especially helpful for designing robots that must function in circumstances in which humans have limited intuition (nanoscale, space, etc.). Evolved simulated robots can also be used as scientific tools to generate new hypotheses in biology and cognitive science, as well as to test old hypotheses that require experiments that have proven to be difficult or impossible to carry out in reality. This can be done by both simulating new data and simulating the results of previous experiments.

At the beginning of the 1990s, two distinct European groups each exhibited a different strategy for how the growth of robot control systems may proceed. EPFL's Dario Floreano and Francesco Mondada are responsible for the development of the Khepera robot's controls. On the other hand, the bodies of these robots were assumed to exist before evolution.

Also in the early 1990s, researchers at the MIT Media Lab named Karl Sims and Jeffrey Ventrella were credited with publishing the results of the first simulations of evolving robots.

{End Chapter 1}

Chapter 2: Evolutionary algorithm

An evolutionary algorithm, often known as an EA, is a general population-based metaheuristic optimization algorithm. It is a subset of evolutionary computation, which is a component of computational intelligence (CI). The processes of biological evolution, including as reproduction, mutation, recombination, and selection, are used in an evolutionary algorithm (EA). The candidate solutions to the optimization issue act as