An Efficient Software Test Method for the Autonomous Mobile Robot Control Program

ROS (Robot Operating System) is widely used as an open-source robotics system for various robot development. In particular, the ROS runtime is composed of several loosely coupled processes, and this loosely coupled structure is designed to allow developers to add each functional module flexibly according to the required functions of the robot. However, software testing of it is required. However, the development of a robot involves many disciplines such as mechanics, electronics, communications, and software, and as a systematic project, there is a high threshold for the average developer. For this reason. In this paper, we propose an efficient and reliable software testing method by introducing Docker to containerize the source code and runtime environment, and putting the build process into a continuous integration platform. For the components of the control software, this paper designs a unit testing scheme with functional packages as units; while for the integrated functions, a big-bang integration testing method is used to test the integrated navigation functions, and the control system is tested with the highest efficiency through an automated testing method. Finally, the test design method is carried out for the simulation experiment of ROS-based four-wheel drive autonomous robot to verify the feasibility of this scheme.