以物聯網結合自動化監控技術改善廢容器稽核認證作業模式
| 中文摘要 | 本計畫規劃以IoT的概念導入感測裝置、行動裝置等工具，並強化業者自我管理及稽核認證效益。工作內容包含：運用AI偵測技術推動數據匯入、導入IoT工具執行智慧監控管理、研擬廢容器雜質自主查驗機制。目標在建置自動化監測設備並開發管理模式，以有效精進稽核認證作業模式、減少稽核人力成本支出。 在前兩項工作的執行上，完成車牌辨識系統、地磅資訊即時連線以及處理設備IoT智慧電錶掛表量測等作業。總計比對後車牌辨識率成功約為92%；地磅資訊即時連線之重量數據吻合度為69%；IoT智慧電錶掛表量測共計收集226,012筆資料，比對後可明顯判斷設備實際運轉時間與產能合理性。 在研擬廢容器雜質自主查驗機制方面，嘗試引用KLR訊號法建立預警機制，以取代現行人力稽核方式。本團隊遴定9項與處理量有關的申報資料，再衍生為12項預警變數，並按監控方程式與預警變數進行二種異常情境的模擬。隨後由RRMS系統截取歷年資料，以建立適當之預警訊號門檻值。雖然針對所需業者申報資料於現行RRMS系統中尚未完全建置，但本研究以設立假設條件之方式進行各項申報數據與預警變數之估算，另外針對異常訊號門檻值判定之方式提出修正建議，最後依此方式就異常訊號判定與試行結果進行討論。 | ||
|---|---|---|---|
| 中文關鍵字 | 智慧電錶、物聯網、車牌辨識、預警系統、KLR訊號法 | ||
基本資訊
| 專案計畫編號 | 經費年度 | 110 | 計畫經費 | 4700 千元 | |
|---|---|---|---|---|---|
| 專案開始日期 | 2021/01/01 | 專案結束日期 | 2021/11/30 | 專案主持人 | 林俊旭 |
| 主辦單位 | 回收基管會 | 承辦人 | 林高正 | 執行單位 | 財團法人中華經濟研究院 |
成果下載
| 類型 | 檔名 | 檔案大小 | 說明 |
|---|---|---|---|
| 期末報告 | EPA-110-XA02.pdf | 6MB |
Using IoT with auto-monitoring technology to improve the procedures of audit and certification for waste container recycling
| 英文摘要 | This project plans to introduce the concept of IoT to sensor devices, mobile devices and other tools, and to strengthen the self-management and audit certification benefits of the industry. The work includes: the use of AI detection technology to promote data import, the introduction of IoT tools to perform intelligent monitoring and management, and the development of an independent inspection mechanism for waste container impurities. The goal is to build automated monitoring equipment and develop a management model to effectively refine the audit and certification operation mode and reduce audit labor costs. In the implementation of the first two tasks, we have completed the license plate recognition system, real-time connection of weighbridge information, and IoT smart meter measurement of processing equipment. After the comparison, the success rate of vehicle license plate recognition is 92%; the weight data of the real-time connection of weighbridge information matches 69%; and 226,012 data of IoT smart meter measurement are collected, which can clearly determine the actual operation time and production capacity of the equipment after the comparison. In the study of waste container impurities independent inspection mechanism, we try to establish an early warning mechanism by using the KLR signal method to replace the current human audit method. The team selected 9 reporting data related to the processing volume, and then derived 12 alert variables, and simulated two kinds of abnormal situations according to the monitoring equation and the alert variables. Then, the RRMS system intercepted the previous years' data to establish the appropriate threshold values for the early warning signals. Although the required reporting information is not yet fully established in the current RRMS, this study uses hypothetical conditions to estimate the reporting data and early warning variables, and proposes amendments to the way of determining the threshold values of abnormal signals. | ||
|---|---|---|---|
| 英文關鍵字 | Smart meter, Internet of Things, IoT, Automatic License Plate Recognition, ALPR, Early warning system, KLR Signals Approach | ||