請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/77840
標題: | 溫室飛行機器人環境偵測結合物聯網系統之研發 Development of a Greenhouse Flying Robot System for Environmental Sensing and IoT Applications |
作者: | Chieh-Yu Lin 林劼佑 |
指導教授: | 陳世銘(Suming Chen) |
關鍵字: | 四軸飛行器,室內定位,環境偵測,自動控制,物聯網,溫室, Quadcopter,Indoor Positioning,Environmental Sensing,Automatic Control,Internet of Things (IoT),Greenhouse, |
出版年 : | 2017 |
學位: | 碩士 |
摘要: | 溫室是最適合進行自動化栽培作物的環境之一。一個自動化環控溫室通常會有溫度、濕度及照度感測器並搭配灌溉、通風、降溫、加熱系統對溫室內環境進行一系列的調控,然而若僅用少量感測器代表整座溫室之環境狀態,在環境狀態不均勻的溫室中進行相同的栽培作業,對於作物的收成及生產者的利潤都是不利的。因此,有的研究於溫室內部佈置大量的感測器節點,根據溫室內部各區域進行因地制宜之栽培手段,但卻也同時增加溫室之建設成本且系統也有維護上的成本。
因此,為能夠僅利用少量之感測器即可得到環境內各位置之環境資訊,本研究探討利用飛行器應用於環境資訊之自動化偵測任務之可行性。本研究以四軸飛行器搭載溫度、濕度、光度感測器進行環境資訊之偵測,研發自動化環境偵測機器人系統,並以物聯網之基礎架構建立網路資料庫進行環境數據之蒐集,作為往後進行大數據分析及應用之數據來源。而本研究開發之環境偵測機器人系統可分為五大子系統,分別為環境偵測飛行機器人、超寬頻 (Ultra-Wideband, UWB) 室內定位系統、決策控制系統、環境數據感測系統與物聯網大數據管理系統。室內定位系統以UWB定位晶片為系統核心,搭配DSTWR (Double-Sided Two-Way Ranging) 距離演算法進行距離估算,使用最小平方法取得飛行器位置,並透過決策控制系統控制飛行器於室內環境中進行自動化巡航作業。於欲量測之目標點懸停後透過機載之環境數據感測系統進行環境資訊收集,最後在完成巡航作業後將感測資訊上傳至物聯網大數據管理系統,提供使用者透過網站介面遠端查詢環境內部情形之服務。在DSTWR距離量測驗證中,其距離誤差RMSE (均方根誤差)平均為0.04 m。在座標估算驗證方面,系統輸出位置相對實際位置誤差於X軸、Y軸之RMSE值分別為0.07 m、0.04 m,絕對距離之RMSE值為0.08 m。本研究中飛行器於定點懸停測試時,於六重複的條件下,其距離誤差RMSE之平均值於X、Y、Z軸分別為4.86 cm、6.11 cm、2.22 cm,飛行器座標與目標位置之絕對距離的RMSE值之平均值則為8.47 cm,其代表於飛行器的控制穩定性上已有相當的水準。在巡航測試中,飛行器於室內環境進行15點之自動巡航並於目標點上懸停5秒鐘進行環境資訊收集,一共花費204.39秒。 綜合以上,本研究成功的建立一套環境偵測機器人系統,環境機器人能進行自動巡航作業並同時收集溫度、濕度、光度之環境資訊,最後將感測結果上傳至網路資料庫,並可透過網站介面進行查詢。證實飛行器應用於環境資訊之自動化偵測任務具有可行性。 Greenhouse is one of the most suitable environments for crop automation production. An automated environmental control greenhouse is equipped with sensors of temperature, humidity, lighting and systems of irrigation, ventilation, cooling and heating. In general, the environmental conditions in the greenhouse are not uniform. Cultivation decision and practice based on the information from a single or just a few sensors to represent the environmental conditions in the whole greenhouse will cause the loss of crop yield and profits. As a result, some past studies, in which a large number of sensor nodes were installed in the greenhouse to obtain more environmental conditions in line with the local conditions, has been reported. However, this approach will lead to the cost increase of construction and maintenance. In order to acquire the site specific environmental information in high resolution with only a few sensors, this research explores the feasibility of flying robots to conduct automatic detection. In this study, a quadcopter equipped with temperature, humidity and lighting quantum sensors has been developed to acquire the environmental conditions. In addition, in view of developing big data analysis and the advanced applications in the future, this study constructed an Internet database on IoT (Internet of Things) structure to collect environmental information. This system could be divided into two main parts with five subsystems including: 1) flying robotic system, 2) UWB (Ultra-Wideband) indoor positioning system, 3) decision control system, 4) environmental information sensing system and 5) IOT data management system. UWB indoor positioning system adopted “UWB positioning chip” as the core, and used DSTWR (Double-Sided Two Way Ranging) algorithm and least squares indoor positioning algorithm to estimate the distance and the position of the flying robot respectively. Decision control system guided the flying robot to automatically cruise under indoor environment by using automatic control theory. Environmental information sensing system collected data when the robot flies close to the target point and then uploaded the data to the IOT data management system after finished the cruise mission to provide remote searching service. According to the distance testing with indoor positioning system (IPS), the RMSE (Root Mean Square Error) value is 0.04 m. In the positioning testing of IPS, the RMSE values of X, Y-axis are 0.07 m, 0.04 m, and the RMSE value of absolute distance is 0.08 m. In the positioning of the flying robot, the RMSE values of X, Y, Z-axis appears to be 4.86 cm, 6.11 cm, and 2.22 cm; and the RMSE value of absolute distance has shown 8.47 cm after six repetitive tests. The results of positioning indicate that the precision of control has reached to a good outcome. In terms of operation time, it requires 204.39 seconds to accomplish a 15-target-point cruise mission, meaning a 5-second hovering time for each target point. In summary, this study has successfully developed a flying robot system for environmental sensing, which can automatically cruise in an indoor environment, simultaneously collect temperature, humidity, lighting quantum data, upload these data to the Internet database and provide remote searching service accordingly. This study also proved the feasibility of using a flying robot to conduct the automatic indoor environment sensing missions. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/77840 |
DOI: | 10.6342/NTU201703847 |
全文授權: | 有償授權 |
顯示於系所單位: | 生物機電工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-106-R04631002-1.pdf 目前未授權公開取用 | 5.81 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。