MODELING OF EMBEDDED FUZZY SYSTEM TO CONTROL HYDROPONIC CROPS

Resumo

The hydroponics consists of a technique for cultivation that dispenses the use of soil, cul-
tivating the plants in water medium, from a solution enriched with nutritious substances necessary for

the development of the plant. Although still incipient, there is a significant increase in the use of this
technique in several locations where the unstable climate harms agricultural production. In this sense,
automation is extremely necessary to minimize the losses, through the control of several variables in
crop shelters. In addition to minimizing the risk of crop failure, automation enables a higher quality

of products by providing optimal conditions for production, allowing autonomy in cultivation and, con-
sequently, that the human workforce is directed to activities that require greater dedication. This paper

presents a system of multiple inputs and multiple outputs (MIMO) that allowing to control the production

and the monitoring of vegetables in greenhouses for cultivation, collecting and controlling essential vari-
ables such as temperature, humidity, luminosity, among others, using fuzzy logic. This system optimizes

production by automating data collection and hydroponic crop management in order to minimize errors

resulting from these activities. First, the study of the variables and the survey of the components neces-
sary for the assembly of the system were realized, defining the following input variables: temperature,

humidity, time and luminosity as input variables, and LEDs irradiation and ventilation, as output vari-
ables. Then, a study of how these variables are related in the fuzzy system decision-making process was

carried out. The Mamdani model was chosen as an inference method, since it is better suited to use in
systems to support decision making. Thus, the fuzzy sets and rules of the Mamdani model were defined.
The fuzzy controller was implemented in C language using an Arduino Mega 2560 board. In order to
verify the operation of the system, tests were performed in the data collection stage, inference procedure
and output activation. The test results converged, and the system outputs responded appropriately.