Hands-On ROS for Robotics Programming

Program highly autonomous and AI-capable mobile robots powered by ROS

Bernardo Ronquillo Japon 지음

Packt(GCO Science) 2020년 02월 26일 출간

작품소개

책소개

▶Book Description
Connecting a physical robot to a robot simulation using the Robot Operating System (ROS) infrastructure is one of the most common challenges faced by ROS engineers. With this book, you'll learn how to simulate a robot in a virtual environment and achieve desired behavior in equivalent real-world scenarios.

This book starts with an introduction to GoPiGo3 and the sensors and actuators with which it is equipped. You'll then work with GoPiGo3's digital twin by creating a 3D model from scratch and running a simulation in ROS using Gazebo. Next, the book will show you how to use GoPiGo3 to build and run an autonomous mobile robot that is aware of its surroundings. Finally, you'll find out how a robot can learn tasks that have not been programmed in the code but are acquired by observing its environment. You'll even cover topics such as deep learning and reinforcement learning.

By the end of this robot programming book, you'll be well-versed with the basics of building specific-purpose applications in robotics and developing highly intelligent autonomous robots from scratch.

▶What You Will Learn
- Get to grips with developing environment-aware robots
- Gain insights into how your robots will react in physical environments
- Break down a desired behavior into a chain of robot actions
- Relate data from sensors with context to produce adaptive responses
- Apply reinforcement learning to allow your robot to learn by trial and error
- Implement deep learning to enable your robot to recognize its surroundings

▶Key Features
- Learn fundamental ROS concepts and apply them to solve navigation tasks
- Work with single board computers to program smart behavior in mobile robots
- Understand how specific characteristics of the physical environment influence your robot's performance

▶Who This Book Is For
If you are an engineer looking to build AI-powered robots using the ROS framework, this book is for you. Robotics enthusiasts and hobbyists who want to develop their own ROS robotics projects will also find this book useful. Knowledge of Python and/or C++ programming and familiarity with single board computers such as Raspberry Pi is necessary to get the most out of this book.

목차

▶TABLE of CONTENTS
？ Section 1: Physical Robot Assembly and Testing
1. Assembling the Robot
2. Unit Testing of GoPiGo3
3. Getting Started with ROS
？ Section 2: Robot Simulation with Gazebo
4. Creating the Virtual Two-Wheeled ROS Robot
5. Simulating Robot Behavior with Gazebo
？ Section 3: Autonomous Navigation Using SLAM
6. Programming in ROS - Commands and Tools
7. Robot Control and Simulation
8. Virtual SLAM and Navigation Using Gazebo
9. SLAM for Robot Navigation
？ Section 4: Adaptive Robot Behavior Using Machine Learning
10. Applying Machine Learning in Robotics
11. Machine Learning with OpenAI Gym
12. Achieve a Goal through Reinforcement Learning

출판사 서평

▶ Preface
Why a new book about learning robotics with ROS? Well, programming is but a small part of what it takes to work with robots. If you want to become really good at robotics, you'll need skills in other areas as well: electromechanics, robot simulation, autonomous navigation, and machine learning/reinforcement learning. Each of these four topics is a building block that you will need to master on your path to acquiring full robotics skills. This book is divided into four parts, each one being devoted to each of these building blocks.

？ Section 1, Physical Robot Assembly and Testing, focuses on electromechanics and describes each hardware part of the robot, providing practical demonstrations of how to test every sensor and actuator that it is equipped with. This part of the book should provide you with a good understanding of how a mobile robot works.

？ Section 2, Robot Simulation with Gazebo, deals with robot simulation. It is here where we introduce ROS and develop a two-wheeled robot simulation that emulates both the physical aspects and the behavior of an actual robot. We will explore the concept of the digital twin, a virtual robot that is the twin of a physical one. This is a fundamental part of developing robotic applications, as it cuts the costs associated with testing real hardware. The digital twin allows us to speed up the development process and save testing with the physical robot for the advanced stages of development.

？ Section 3, Autonomous Navigation Using SLAM, is devoted to robot navigation, the most common task for mobile robots. State-of-the-art algorithms and techniques are explained in a practical manner, first in simulation and then with a physical robot.

？ Section 4, Adaptive Robot Behavior Using Machine Learning, focuses on machine learning and reinforcement learning, the most active fields in robot research and real-world robotic applications. By using this technology, a robot is able to transition from pure automatism ？ where every possible behavior or answer is coded ？ to being a flexible behavior machine, where the robot is capable of reacting in a smart way to environmental demands by learning from data. This data can be obtained from the robot's previous experience or gathered from the experience of similar robots.

To build a state-of-the-art robot application, you will first need to master and then combine these four building blocks. The result will be what is commonly known as a smart robot. This is your task ？ this is your challenge.

인물정보

저자(글) Bernardo Ronquillo Japon