人工智慧與物聯網技術在近年來崛起，藉由所佈建的感測裝置，與所收集到的巨量感測數據，使得數據感知分析成為熱門研究議題。本研究基於人工智慧與物聯網等技術並結合農業專家知識，研發農業專家栽培系統，進行無人化的栽培管理與農場維運，目的在於建立最佳化的灌溉排程與自動化的栽培系統。
本研究建構一Artificial Intelligence of Things (AIoT)架構，在感測層中使用低功耗的室外物聯網感測裝置，本研究在感測器資料傳輸上提出Dynamic Environment Detection 機制，其為以數據特性進行資料傳輸，能降低整體的傳輸次數，達到節能的效益。通訊層採用低功耗 LoRaWAN 搭配Message Queuing Telemetry Transport (MQTT)協議，來提升傳輸距離與降低能源消耗。邊緣層包含AI module與Irrigation module，在AI module中，資料收集模組 (Data Collect Function)進行感測資料收集。生長狀態模組(Growth Status Function)進行作物生長狀態的分析，利用色塊執行作物成熟度的計算。植物需水量模組(Plant Water Requirement Function)用來預測作物需水量，其中使用Extreme Gradient Boosting (XGBoost)建立預測模型，XGBoost的演算架構有效降低Variance並減少Bias，能夠實地貼合真實的應用情境。澆灌決策模組(Watering Decision Function)依據專家經驗做澆灌排程並執行澆灌預測。Irrigation module負責接AI module的訊息並下達到澆灌控制器中，以達無人化栽培。
本研究分別探討GAM(Generalized Additive Model)、SVM(Support Vector Machine)、Random Forest與XGBoost的預測效果，其中以XGBoost的預測效果最佳，MSE值為0.0085最低。實驗分析比較本研究所提的架構、Fixed與Threshold的澆灌時間，實驗結果發現本研究相較於Fixed與Threshold分別少了50%與30%的用水量，能夠更精準的實現灌溉用水。在Dynamic environment detection的傳輸機制，相比定時的傳輸模式，減少97.8%的通訊次數，有效降低整體的傳輸能耗。在人力資源分析中，比較本研究與農夫實地栽培的時間，實驗結果發現本研究能夠節省40.94%的栽培時間，並且更有效的進行農場的管理。由上述實驗結果得知，本研究提出之方法，能夠提供栽培者一個便利的栽培模式。

Artificial intelligence (AI) and the Internet of Things (IoT) technologies have been on the rise in recent years. These technologies are based on installed sensors and big data that these sensors collect. Data sensing and analysis have therefore become a popular research topic. The present study integrated AI and the IoT with the knowledge of agricultural experts to develop an expert agricultural cultivation system that realizes cultivation management and farm operations without human labor for the purpose of establishing an optimal irrigation schedule and automated cultivation system.
In the present study, an Artificial Intelligence of Things architect that uses low-power-consumption outdoor IoT sensors in the sensing layer was constructed. A dynamic environment detection mechanism was proposed for sensor data transmission, in which data was transmitted on the basis of data characteristics to reduce the total number of transmission times and save energy. In the communication layer, low-power-consumption Long Range Wide Area Network (LoRaWAN) was used in combination with Message Queuing Telemetry Transport protocol to increase the transmission distance and reduce energy consumption. The edge layer consisted of an AI module and irrigation module. In the AI module, a data collection function was used to collect sensing data, and a growth status function was used to calculate the maturity level of crops by using color blocks. A plant water requirement function was used to predict the amount of water needed by the crops. Extreme gradient boosting (XGBoost) was adopted to build a prediction model. The algorithm of XGBoost effectively reduces variance and reduce bias, thereby ensuring that the prediction reflects the actual situations. The watering decision function schedules watering according to expert experience and performs watering prediction. The irrigation module receives messages from the AI module and delivers them to the watering controller, thereby completing a cultivation system without human labor.
This study compared the prediction accuracy of the generalized additive model, support vector machine, random forest, and XGBoost and determined that XGBoost had optimal prediction accuracy and obtained the lowest mean square error (0.0085). The experimental results of the proposed architecture were revealed to show 50% and 30% less water consumption during irrigation than those obtained from using fixed and threshold watering times, respectively. The proposed method demonstrates more precise watering of crops. The dynamic environment detection transmission mechanism exhibited a 97.8% lower communication frequency compared with the fixed-schedule transmission mode and effectively reduced the overall energy consumption during transmission. In the analysis of human labor, the proposed system was discovered to reduce the cultivation time by 40.94% compared with manual cultivation, thus facilitating more effective farm management. The experimental results indicated that the proposed method can serve as a convenient cultivation model for farmers.

[1]Fengjiao Jia and Yisong Li, “Resident's consumer demand, government assignment and development of urban agricultural products logistics — Based on the data of agricultural products market in Beijing,” Proceedings of International Conference on Logistics, Informatics and Service Sciences, pp. 1-4, 2016.
[2]Guang Li and Shuang Yu, “Study on construction of agricultural product supply chain and benefit distribution based on agricultural industrialization poverty relief project: Taking Rosa sterilis processing industry as the example,” Proceedings of International Conference on Service Systems and Service Management, pp. 1-3, 2017.
[3]Xu Chen and Hua Fang, “The analysis of agricultural products consumers' purchase behavior under the background of big data,” Proceedings of International Symposium on Computer, Consumer and Control, pp. 420-423, 2018.
[4]Yingfeng Ji, Hualong Yang and Meitong Chen, “Logistics network configuration for fresh agricultural products,” Proceedings of 29th Chinese Control And Decision Conference, pp. 5724-5727, 2017.
[5]Chenggong Li, Yongwei Tang, Maoli Wang and Xiaojie Zhao, “Agricultural machinery information collection and operation based on data platform,” Proceedings of IEEE International Conference of Safety Produce Informatization, pp. 472-475, 2018.
[6]Ji-chun Zhao and Jian-xin Guo, “Big data analysis technology application in agricultural intelligence decision system,” Proceedings of IEEE 3rd International Conference on Cloud Computing and Big Data Analysis, pp. 209-212, 2018.
[7]Zhaochan Li, Jinlong Wang, Russell Higgs, Li Zhou and Wenbin Yuan, “Design of an intelligent management system for agricultural greenhouses based on the internet of things,” Proceedings of IEEE International Conference on Computational Science and Engineering and IEEE International Conference on Embedded and Ubiquitous Computing, pp. 154-160, 2017.
[8]Mukesh Kumar Tripathi and Dhananjay D. Maktedar, “Recent machine learning based approaches for disease detection and classification of agricultural products,” Proceedings of International Conference on Computing Communication Control and automation, pp. 1-6, 2016.
[9]Hu Haiyan and Chen Tao, “Design and implementation of agricultural production and market information recommendation system based on cloud computing,” Proceedings of 8th International Conference on Intelligent Computation Technology and Automation, pp. 367-370, 2015.
[10]Zhang Rong and Liu Bin, “Farmer or logistics agent: Who should launch the RFID in an agricultural product supply chain?,” Proceedings of 12th International Conference on Service Systems and Service Management, pp. 1-5, 2015.
[11]Danping Lin, C. K. M. Lee and W. C. Tai, “Application of interpretive structural modelling for analyzing the factors of IoT adoption on supply chains in the Chinese agricultural industry,” Proceedings of IEEE International Conference on Industrial Engineering and Engineering Management, pp. 1347-1351, 2017.
[12]Konlakorn Wongpatikaseree, Promprasit Kanka and Arunee Ratikan, “Developing smart farm and traceability system for agricultural products using IoT technology,” Proceedings of IEEE/ACIS 17th International Conference on Computer and Information Science, pp. 180-184, 2018.
[13]Md Shadman Tajwar Haque, Khaza Abdur Rouf, Zobair Ahmed Khan, Al Emran and Md. Saniat Rahman, “Design and implementation of an IoT based automated agricultural monitoring and control system,” Proceedings of International Conference on Robotics,Electrical and Signal Processing Techniques, pp. 13-16, 2019.
[14]Ramya Venkatesan and Anandhi Tamilvanan, “A sustainable agricultural system using IoT,” Proceedings of International Conference on Communication and Signal Processing, pp. 763-767, 2017.
[15]S. Rajeswari, K. Suthendran and K. Rajakumar, “A smart agricultural model by integrating IoT, mobile and cloud-based big data analytics,” Proceedings of International Conference on Intelligent Computing and Control, pp. 1-5, 2017.
[16]Donghui Wu, “A big data analytics framework for forecasting rare customer complaints: A use case of predicting MA members' complaints to CMS,” Proceedings of IEEE International Conference on Big Data, pp. 3965-3967, 2017.
[17]Antonino Galletta, Lorenzo Carnevale, Alina Buzachis, Antonio Celesti and Massimo Villari, “A microservices-based platform for efficiently managing oceanographic data,” Proceedings of 4th International Conference on Big Data Innovations and Applications, pp. 25-29, 2018.
[18]Bi Puyun and Li Miao, “Research on analysis system of city price based on big data,” Proceedings of IEEE International Conference on Big Data Analysis, pp. 1-4, 2016.
[19]Li Yang and Jun-Jie Zhang, “Realistic plight of enterprise decision-making management under big data background and coping strategies,” Proceedings of IEEE 2nd International Conference on Big Data Analysis, pp. 402-405, 2017.
[20]Qingwu Hu and Yuan Zhang, “An effective selecting approach for social media big data analysis — Taking commercial hotspot exploration with Weibo check-in data as an example,” Proceedings of IEEE 3rd International Conference on Big Data Analysis, pp. 28-32, 2018.
[21]Sk Al Zaminur Rahman, Kaushik Chandra Mitra and S.M. Mohidul Islam, “Soil classification using machine learning methods and crop suggestion based on soil series,” Proceedings of 21st International Conference of Computer and Information Technology, pp. 1-4, 2018.
[22]Laxmikant Bordekar, Harish Velingkar, Elanton Fernandes, Heramb Hanumant Bandekar, Ashutosh Gurudas Harmalkar and Bradwel Jose Antonio Pinto, “Cashew nut grade identification and quality testing using machine learning,” Proceedings of Second International Conference on Inventive Communication and Computational Technologies, pp. 661-664, 2018.
[23]Parth Mehta, Hetasha Shah, Vineet Kori, Vivek Vikani, Soumya Shukla and Mihir Shenoy, “Survey of unsupervised machine learning algorithms on precision agricultural data,” Proceedings of International Conference on Innovations in Information, Embedded and Communication Systems, pp. 1-8, 2015.
[24]Hardik B. Sailor, Hemant A. Patil and Avni Rajpal, “Unsupervised filterbank learning for speech-based access system for agricultural commodity,” Proceedings of Ninth International Conference on Advances in Pattern Recognition, pp. 1-6, 2017.
[25]Chandrasegar Thirumalai, K Sri Harsha, M Lakshmi Deepak and K Chaitanya Krishna, “Heuristic prediction of rainfall using machine learning techniques,” Proceedings of International Conference on Trends in Electronics and Informatics, pp. 1114-1117, 2017.
[26]Randy Erfa Saputra, Budhi Irawan and Yakub Eka Nugraha, “System design and implementation automation system of expert system on hydroponics nutrients control using forward chaining method,” Proceedings of IEEE Asia Pacific Conference on Wireless and Mobile, pp. 41-46, 2017.
[27]Maryam Hazman, “Crop irrigation schedule expert system,” Proceedings of 13th International Conference on ICT and Knowledge Engineering, pp. 78-83, 2015.
[28]Karolina Szturo and Piotr M. Szczypiński, “Ontology based expert system for Barley grain classification,” Proceedings of Signal Processing: Algorithms, Architectures, Arrangements, and Applications, pp. 360-364, 2017.
[29]Gayatri Dwi Santika, Diah Ayu Retnani Wulandari and Fitriyana Dewi, “Quality assessment level of quality of cocoa beans export quality using hybrid Adaptive Neuro - Fuzzy Inference System (ANFIS) and Genetic Algorithm,” Proceedings of International Conference on Electrical Engineering and Computer Science, pp. 195-200, 2018.
[30]Mostafa Sharifi and XiaoQi Chen, “A novel vision based row guidance approach for navigation of agricultural mobile robots in orchards,” Proceedings of 6th International Conference on Automation, Robotics and Applications, pp. 251-255, 2015.
[31]Priyanka Soni and Rekha Chahar, “A segmentation improved robust PNN model for disease identification in different leaf images,” Proceedings of IEEE 1st International Conference on Power Electronics, Intelligent Control and Energy Systems, pp. 1-5, 2016.
[32]Hülya Yalçın, “Phenology monitoring of agricultural plants using texture analysis,” Proceedings of 24th Signal Processing and Communication Application Conference, pp. 1061-1064, 2016.
[33]Yi Zhao and Xiaohui Li, “Edge-aware weighting enhanced saliency segmentation of pests images,” Proceedings of International Conference on Computational Science and Computational Intelligence, pp. 847-851, 2016.
[34]G. Kavianand, V.M. Nivas and R. Kiruthika, “Smart drip irrigation system for sustainable agriculture,” Proceedings of IEEE Technological Innovations in ICT for Agriculture and Rural Development, pp.19-22, 2016.
[35]J. Zhao and J. Guo, “Big data analysis technology application in agricultural intelligence decision system,” Proceedings of IEEE 3rd International Conference on Cloud Computing and Big Data Analysis , pp. 209-212 , 2018.
[36]Lijia Sun, Yanxiang Yang, Jiang Hu, Dana Porter, Thomas Marek and Charles Hillyer, “Reinforcement learning control for water-efficient agricultural irrigation,” Proceedings of IEEE International Symposium on Parallel and Distributed Processing with Applications and IEEE International Conference on Ubiquitous Computing and Communications, pp.1334-1341, 2017.
[37]Ban Alomar and Azmi Alazzam, “A smart irrigation system using IoT and fuzzy logic controller,” Proceedings of Fifth HCT Information Technology Trends, pp. 175-179, 2018.
[38]M. Juan, V. Núñez, R.F. Fonthal and M.Q.L Yasmín, “Design and implementation of WSN for precision agriculture in white cabbage crops,” Proceedings of IEEE XXIV International Conference on Electronics, Electrical Engineering and Computing, pp. 1-4, 2017.
[39]K.A. Patil and N.R. Kale, “A model for smart agriculture using IoT,” Proceedings of International Conference on Global Trends in Signal Processings of the Information Computing and Communication, pp.543-545, 2016.
[40]W. Dongjie , L. Zhemin and W. Shengwei ,“Exploring the relationship between maize yield and climate big data on maize belt of northeast China,” Proceedings of IEEE 2nd International Conference on Big Data Analysis, pp.513-516 , 2017.
[41]P. Zhang , Q. Zhang and F. Liu ,“The construction of the integration of water and fertilizer smart water saving irrigation system based on big data,” Proceedings of IEEE International Conference on Computational Science and Engineering and IEEE International Conference on Embedded and Ubiquitous Computing, pp.392-397 , 2017.
[42]D. Smith and W. Peng, “Machine learning approaches for soil classification in a multi-agent deficit irrigation control system,” Proceedings of IEEE International Conference on Industrial Technology, pp. 1-6, 2009.
[43]S. Heble, A. Kumar and K.V.V. Prasad, “A low power IoT network for smart agriculture,” Proceedings of IEEE 4th World Forum on Internet of Things, pp.609-614, 2018.
[44]C. Trongtorkid and P. Pramokchon, “Expert system for diagnosis mango diseases using leaf symptoms analysis,” Proceedings of International Conference on Digital Arts, Media and Technology, pp.59-64, 2018.
[45]E. Agustina, I. Pratomo and A.D. Wibawa, “Expert system for diagnosis pests and diseases of the rice plant using forward chaining and certainty factor method,” Proceedings of International Seminar on Intelligent Technology and Its Applications, pp.266-270, 2017.
[46]Y. Liu, X. Zhao and X. Zhu, “Knowledge expression and reasoning model for tomato disease diagnosis,” Proceedings of 3rd International Conference on Information Management, pp.284-288, 2017.
[47]J. Ren, Y. Guo, D. Zhang, Q. Liua and Y. Zhang, “Distributed and efficient object detection in edge computing: challenges and solutions,” Proceedings of IEEE Network, pp. 1-7, 2018.
[48]K. He, X. Zhang, S. Ren and J. Sun, “Deep residual learning for image recognition,” Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 770-778, 2016.