隨著機器人的快速發展及復雜化，代碼復用性需求也越來越強烈，因此 誕生了很多開源機器人系統，ROS（Robot Operating System）就是其中之一。 ROS 是專為機器人軟件開發所設計出來的一套電腦操作系統架構。它是一 個開源的元級操作系統（後操作系統），其中包括硬件抽象描述、底層驅動 程序管理、共用功能的執行、程序間消息傳遞、程序發行包管理，它也提供 一些工具和開源庫用於獲取、建立、編寫和執行多機融合的程序。由於機器 人在自動化產業需求日益趨增，要能投入醫療及加工廠等不同應用場景，行 動過程就必須仰賴環境空間進階感測搭配自身載具導航至定點、同時判斷 被服務對象與機器人間的距離與環境空間限制，本論文主要為完成一套基 於 ROS 系統下運用 SLAM 技術獲取地圖資訊及定位系統，並且加入物件辨 識人體骨架追蹤技術，此套系統主要分為 ZED2 深度攝影機與光達兩種傳 感器，ZED2 主要提供 IMU、RGB 色彩資訊及地圖美化，光達則提供較精 確定位的系統;With the development and specialization of robots, there is a higher demand for the functions of robots. Therefore, many open source robot systems have been born, and ROS (Robot Operating System) is one of them. ROS has its own computer operating system for robotic software development. It is an open source meta-level operating system (post-operating system) that provides a set of kernel services, including hardware description, related driver management, execution of consumption functions, message transfer between programs, and program release package management. It also provides some tools and open source libraries for obtaining, creating, writing and multi-machine integration programs. As the demand for robot automation continues to grow, and to meet different application scenarios such as medical treatment and processing plants in the industry, the action process must rely on the advanced perception of the environmental space of the own vehicle to navigate to a fixed point. At the same time determine the distance between the service object and the robot. This paper is mainly to complete the acquisition of map information and positioning system and add object recognition human skeleton tracking technology based on the motion SLAM technology under the ROS system. This system is mainly divided into two types of sensors: ZED2 depth camera and Lidar. ZED2 mainly provides IMU, RGB color information and map beautification. LiDAR provides a more accurate positioning system and virtual field.
