AGV是倉儲系統中是不可或缺的角色，但AGV的主流引導模式：有軌引導，和更爲先進的引導模式：無軌引導，皆存在諸多缺點。隨著“工業4.0”時代的來臨，智慧倉儲中AGV的引導模式也開始由有軌引導向無軌引導過渡。為銜接兩種引導模式，且解決新舊兩種引導模式的不足，本研究以MIAT方法論為設計基礎，提出了基於ROS2.0的AGV棧板對接位導航系統，此系統結合了SLAM製圖定位、深度學習物件辨識、及二維碼定位演算法，以獲取棧板的精確位置，並將AGV導航至棧板對接位。本系統將兩種引導模式優勢結合互補，提供穩定且高效的引導效率。通過實驗，本系統在簡單場景中能夠達到94%的成功率，複雜場景中導航也能展現出成功率84%的優異表現，證明本系統能夠很好的完成AGV棧板對接位導航這項任務。本研究提出的系統在工作環境條件允許的情況下，可以很好的完成棧板對接位導航這項任務，同時擁有較强的靈活度及優異的成功率。;AGV is an important role in the intelligent warehousing, but there are many shortcomings in the mainstream guidance mode of AGV: visual track guidance, and the state-of-the-art guidance mode: trackless guidance. With the advent of the "Industry 4.0" era, the guidance mode of AGV in intelligent warehousing has also begun to transition from track guidance to trackless guidance. In order to connect the two guidance modes and solve the shortcomings of the old and new guidance modes, this study uses MIAT methodology is the basis of the design, and a ROS2.0-based AGV pallet docking navigation system is proposed. This system combines SLAM mapping positioning, deep learning object recognition, and two-dimensional code positioning algorithms to obtain the precise position of the pallet, and Navigate the AGV to the pallet docking position.This system combines the advantages of the two guiding modes and complements each other to provide stable and efficient guiding efficiency. Through experiments, this system can achieve a success rate of 94% in simple scenes, and the navigation in complex scenes can also show an excellent performance with a success rate of 84%, which proves that this system can well complete the task of navigating the docking position of AGV pallets .The system proposed in this study is capable of navigating the docking position of pallets when the working environment conditions permit, and has strong flexibility and excellent success rate.
