Desarrollo de un sistema de monitoreo para la determinación de nutrientes del suelo presentes en cultivos de cacao basado en smart grid y machine learning.

Abstract

The objective of the Curricular integration research project was to develop a monitoring system based on a Smart grid and machine learning for the determination of soil nutrients present in cocoa crops located at the ESPOCH in Orellana province in La Belleza farm. The correct development of the plants is due to the existing nutrients in the soil, so it was necessary to analyze the macronutrients, nitrogen, phosphorus, and potassium, with a soil NPK fertility sensor. The monitoring system is based on a Smart grid of two nodes, which have Lora Wan technology; it has two Lora drivers, a Gateway UG-65 and a Router, which allows sending the values to the internet using the IoT with storage of data in the cloud. The data is processed through a machine learning algorithm based on fuzzy logic that provides alarms and recommendations regarding the fertilizers that should be used; using the Node-Red software, the information collected in real-time is displayed in the application IoT Cloud and Dashboard programmed in Node-Red. The monitoring system is working correctly, for this several tests were carried out, starting with the correct use of the sensor; it was confirmed that the values shown directly on the computer through the Modbus Poll software, the Node-Red dashboard, and the simulation of the MATLAB algorithm, have a similar mean of the values obtained in the laboratory analysis with an error of less than 2%. In addition, with the normality and T-test, it was concluded that the null hypothesis applied to each nutrient is accepted since the average is greater than the significance level of 0.01.