開發具視覺系統之環保服務機器人應用於封閉型社區

Development of Vision-Based Environmental Management Robots for Gated Communities

本研究在於利用已知的各項智慧技術研究，再結合物聯網技術加以整合，而應用於社區之環保服務型機器人，有幾個技術層面需要探討與執行，分別是機構設計、導航系統、物件識別與物聯控制。由於一般垃圾清運車通常難以進入這類型社區，因此社區會準備各式垃圾與資源回收之組合桶體，以利住戶丟棄，惟仍有少數缺乏公德心的人，不依據回收物屬性進行分類，而造成後續分類回收處理的困擾，本研究所設計之機器人任務將進行二階段的垃圾丟棄，第一階段會視垃圾的類別，如廚餘、有害、其它及回收進行初步分類，若垃圾判斷為前三者，機器人將分別輸送至所屬的垃圾桶，若為回收類別，則運至專屬的回收輸送機上進行第二階段的細項（鐵鋁罐、寶特瓶、鋁箔包及其它等）分類。綜觀目前市場上應用在社區型之環保服務型機器人，大都是具有分類功能但是固定式、或者可移動但只能固定路線巡迴清理垃圾且不分類，本研究希望本著彈性應用（多用途）、低成本與可商品化為出發點設計機器人，此機器人任務未來可再延伸至智慧巡邏、貨物搬運、園藝灌溉、環境維護等功能，進行全面性的社區服務，市場上則可針對接受度高的中豪宅社區進行導入，除協助提升房產形象，亦可創造房價優勢。

This study integrated existing intelligent technologies with the Internet of Things (IoT) to develop an environmental management robot for gated communities. The development of the robot system was focused on mechanical design, visual navigation, object recognition, and IoT-based remote control. Communities typically provide various types of universal bins for general waste and recyclables. However, some residents may dispose of waste improperly in violation of recycling rules, complicating downstream classification and recycling. To address this problem, we developed the aforementioned robot to dispose of waste in two stages. In the first stage, the robot identifies the type of waste. If the waste belongs to any of the first three categories, it is routed to its designated garbage bin accordingly. The second stage involves sorting waste on a dedicated conveyor belt, where materials such as metal cans, PET bottles, and Tetra Pak cartons are further classified. This work presents a modular and scalable robotic platform with low cost, functional flexibility, and high commercialization potential. Compared with current market solutions - most of which are either stationary or constrained to fixed-route cleaning without sorting capabilities - this system offers enhanced flexibility, multi-functionality, and strong commercial potential.