A Beginner’s Guide to ROS Development

Here it comes: GuYueJu has analyzed the job requirements of companies in Beijing, Shanghai, Jiangsu, and other places for those interested in or currently learning this career path, providing a clear, comprehensive, and systematic path for becoming a ROS development engineer. We hope that after reading, you will have a clearer understanding of the learning path for this profession and take the easiest way to become a ROS development engineer.

Figure 1 Overview of the Learning Path

Python:

https://www.bilibili.com/video/BV1c4411e77t?from=search&seid=11983732368546479265

ROS development engineers are responsible for the software system development, integration, and code testing of products based on the ROS system (such as AGV vehicles). This requires familiarity with embedded Linux operating systems, a certain understanding of data structure algorithms, network programming, process communication, and multithreading programming. For the cultivation of software development capabilities, we provide the following learning path and related learning resources:

⑴ Data Structure Algorithms

Figure 2 “Linux and Qt Programming”

Through the first two steps of learning, you have acquired the ability to perform software development in a Linux environment. However, the software system developed is just a framework, and next we need to fill in the application-related content within the framework to gradually improve its functionality. This requires us to master the general theories and technologies of robotics and understand the basic applications of ROS development.

ROS development engineers need to complete the driving of various sensors and actuators under the ROS system, as well as data collection and processing. This requires mastering laser sensors, inertial sensors, as well as odometry, vision, ultrasonic, GPS, and various sensor data fusion processing and tuning methods.

After mastering the principles of common sensors, you can formally start the most critical content—the study of the Robot Operating System (ROS). Here we recommend “21 Lectures on ROS,” which uses the classic turtle case to explain the principles and cases of ROS in an easy-to-understand manner, making it easier to get started with ROS.

“21 Lectures on ROS”

The general applications are mainly divided into two major blocks: perception localization and path planning.

The mainstream localization technologies include ultrasonic navigation positioning, visual navigation positioning, GPS positioning, SLAM, etc. Currently, the mainstream robot localization technology is the SLAM technology (Simultaneous Localization and Mapping). For learning SLAM, we suggest: first familiarize yourself with the commonly used SLAM packages in ROS, and then you can theoretically study and practice the most commonly used VSLAM and Lidar SLAM technologies in robot mapping and navigation within ROS.

“Guide to Robot Path Planning Algorithms Based on Grid Maps”

Mastering an efficient algorithm development process is an essential quality for becoming an excellent engineer. “MATLAB + ROS Rapid Development of Control Algorithms” demonstrates the process of quickly developing control algorithms through a trajectory tracking control algorithm development example—V-shaped development process. This technical route allows for rapid verification of one’s ideas and the accurate and quick generation of ROS C++ code to complete the control system development.

Autonomous navigation is a function that mobile robots generally need to possess. Developing autonomous navigation functions requires understanding the ROS Navigation framework and its core function packages, configuring and implementing the autonomous navigation function of mobile robots in a simulation environment, and using SLAM to achieve simultaneous mapping and autonomous navigation, making your robot smarter.

Multi-joint robotic arms are an important component of industry, capable of executing various tasks and operations with precision. Multi-joint robotic arms are widely used in industrial and social life, such as machine loading, pick-and-place operations, welding, painting, inspection, assembly, mechanical manufacturing, and medical applications. Many companies are also recruiting talents for robotic arm development based on ROS, but the development of robotic arms has a relatively high threshold, and the theoretical part is difficult to get started. The recommended learning path is as follows:

Robot kinematics is the main kinematic algorithm in controllers. Through robot kinematics, robots can accurately reach target points to perform various tasks and operations. Here we recommend a course that leads everyone to learn robot kinematics, master the derivation methods of forward kinematics, learn simulation of forward kinematics and methods for verifying the results of forward kinematics, as well as the basic concepts of robotics: position, posture, coordinate system description, mapping, and transformation.

The goal of studying robotics is to better control robots to achieve target functions. Given that the theory of robotic arm control is profound and requires a good mathematical foundation, GuYueJu recommends combining practical experience with theory for beginners in robotic arm control. In the course “Combining Software and Hardware to Play with ROS Robotic Arms,” students can learn about various aspects of this design project, how to build this robotic arm, how to use Lichuang EDA to draw schematics and PCBs, and how to develop a functional package under ROS.

Drones are the mainstream R&D products of many robotics companies. To ensure the safe operation of drones, simulation is crucial in the ROS development process of drones. Flight control development involves a wide range of knowledge and has a high threshold. Students interested in related positions need to be familiar with:

How to get started with drone path planning/visual algorithms is also a challenge for many beginners. Through the “Drone Simulation and Development Based on ROS” course, you will learn the complete code framework for PX4 drone 3D simulation development without real drone conditions, including building a simulation environment for intelligent drone development, initial autonomous flight of drones, and autonomous flight with laser radar obstacle avoidance, facilitating further intelligent development involving path planning and autonomous decision-making. On this basis, you can implement reinforcement learning, visual tracking guidance, laser/visual SLAM, and other R&D work. This course also provides some demo examples with complete interfaces.

For systematic learning of flight control development, we recommend the classic textbook widely used in universities – “Embedded Flight Control Development Guide for Multirotor UAVs”. This book mainly focuses on the design of flight control systems for multirotor drones, starting from the basics of embedded systems, and introduces basic knowledge of drones and hardware composition in an easy-to-understand manner. It emphasizes the principles, foundations, and development processes of drone flight control systems, introduces several different calculation methods for the state estimation of flight vehicle systems, and provides corresponding practical code examples. This book elaborates on the design of drone systems from various aspects, providing cutting-edge knowledge and information that includes foundational knowledge for beginners and algorithm analysis for industry researchers, as well as principles of indoor positioning SLAM.

Figure 3 “Embedded Flight Control Development Guide for Multirotor UAVs”