新冠肺炎 (COVID-19) 疫情之後，各行業頻繁出現人力短缺問題。許多追求未來技術的工廠正在積極發展智慧工廠，引進自動化設備，以提高工廠製造效率。然而，現有無線通信的延遲和不可靠性使其難以滿足AGV導航的需求。選擇正確的傳感器、可靠的通信和導航控制技術對於系統整合商來說仍然是一個具有挑戰性的問題。當今大多數無人駕駛車輛都使用昂貴的感測器或需要布建新的基礎設施或預先構建工廠電子地圖，這阻礙了其廣泛採用度。在本文中，我們介紹了一種用於無人駕駛車輛系統的自學習路徑規劃和高效率圖像識別演算法。我們開發了一種無需添加任何專門基礎設施即可進行導航的無人駕駛車輛系統，並在工廠進行了測試以驗證其可用性。該系統的創新之處在於，我們開發了一種無需任何額外基礎設施或預建工廠電子地圖的無人駕駛車輛系統，並且我們開發了基於邊緣計算和物聯網技術的混合車隊管理系統，以提高駕駛安全性。該系統的核心貢獻在於，我們為無人駕駛車系統開發了快速圖像識別演算法、自學習路徑規劃演算法和混合車隊管理方法，以提高導航安全性。;Post-COVID-19, there are frequent manpower shortages across industries. Many factories pursuing future technologies are actively developing smart factories and introducing automation equipment to improve factory manufacturing efficiency. However, the delay and unreliability of existing wireless communication make it difficult to meet the needs of AGV navigation. Selecting the right sensor, reliable communication, and navigation control technology remains a challenging issue for system integrators. Most of today’s unmanned vehicles use expensive sensors or require new in-frastructure to be deployed, impeding their widespread adoption. In this paper, we have devel-oped a self-learning and efficient image recognition algorithm. We developed an unmanned vehicle system that can navigate without adding any specialized infrastructure, and tested it in the factory to verify its usability. The novelties of this system are that we have developed an unmanned vehicle system without any additional infrastructure, and we developed a rapid image recognition algo-rithm for unmanned vehicle systems to improve navigation safety. The core contribution of this system is that the system can navigate smoothly without expensive sensors and without any ad-ditional infrastructure. It can simultaneously support a large number of unmanned vehicle systems in a factory.
