With the rapid development of robotics, intelligent obstacle avoidance robots are receiving more and more attention. They are heavily used in all aspects of production and development. The framework we chose to develop this time is based on an open-source platform called Robot Operating System (ROS), which acts as a digital twin in a modular configuration. There are now a large number of robots that choose to complete their architecture under the Robot Operating System (ROS), such as disaster relief robots, somatosensory robots, delivery robots, and more. In this article, we will introduce the ROS architecture of our project. This architecture enables robots to map and navigate autonomously in different environments with different sensors. This article will analyze the architecture we have made and explain the choice to make such an architecture. At the same time, we will also analyze our performance and point out which parts can be improved.