Design and implementation of a system of automation, monitoring and control of crops in plastic greenhouses using

The Internet of Things (IoT) has had a very significant development in the last years, mainly because of its wide variety of applications. From connecting a toaster to connecting a series of industrial machines to the Internet translates into greater efficiency and resource savings. The IoT impact has been so strong that even traditional interconnection models, such as OSI or TCP/IP, have been modified to adapt to this new technological revolution. Combining web development techniques with IoT makes it possible to create tools to control devices instantly. One of the most potential fields for applying these integrated tools is precision agriculture. Thus, this work proposes a system that automates and evaluates the environmental variables of greenhouse crops. This IoT system can control and monitor crops through electronic devices and analyze the data acquired by digital and analogical sensors. This system uses a series of devices like sensors, actuators, and communication nodes to measure environmental variables inside a greenhouse, such as environmental temperature, environmental and soil humidity, irrigation, ventilation, and water level. The end-user can make decisions based on the acquired measurements displayed in a graphical interface as a time series. The measurements and user interactions travel through an intermediary application (broker) using ESP8266 NodeMCU modules connected to the Internet by WiFi. The system employs an EMQX broker using the MQTT protocol for data transmission. Given the nature of MQTT implementation and performance, it consumes less bandwidth than the HTTP protocol, usually used for data transfer. The system also consists of a web application that reads, stores, and displays the data in real-time in a graphical user interface (GUI). One of the most critical functions of this application is the broker connection that manages the flow of both read and write signals, i.e., what the user sees and controls. In its final stage, the project uses artificial intelligence tools to obtain relevant predictions of environmental variables based on data collected over time. A recurrent neural network known as Long Short-Term Memory (LSTM) is in charge of the forecast generation by analyzing the time series generated by the data readings.

• 1. Verfasser: Barba Carvajal, Marcelo Leonel (author)
• Format: bachelorThesis
• Sprache: eng
• Veröffentlicht: 2020
• Schlagworte: Internet de las cosas; Invernaderos; Agricultura de precisión; Análisis de datos; Internet of things; Greenhouses; Precision agriculture; Data analysis
• Online Zugang: http://repositorio.yachaytech.edu.ec/handle/123456789/263