Implementation of an AIoT-Based Intelligent Water Resources Control System for Smart Farm

Abstract

This paper presents an artificial intelligence of things (AIoT) based intelligent water resource control (IWRC) system for hydroponic lettuce and fingerroot farms. Three input variables temperature, humidity, and amount of sunlight were used to analyze the optimal value of the output variable with pulse width modulation to control the water pump. The mathematical equations included multiple linear regression (MLR) and adjusting the system with particle swarm optimization and then training the process in the adaptive neuro - fuzzy inference system to generate the Arduino program code. The Arduino board controller can employ the Embedded Fuzzy Logic Library and display values on the Node-RED dashboard. In addition, the proposed system can analyze the performance with mean square error (MSE), root mean square error (RMSE), and R-squared (R2) values. The results show that complex problems in agriculture can be resolved using AIoT by applying intelligent water resource systems in smart farms. All are seamlessly integrated and compatible with Arduino boards, referred to as the MPAEN-Ar algorithm. The most efficient algorithm was the MLR-PSO-ANFIS444 algorithm, which yielded an RMSE of 0.0002345207880 and R2 value of 0.99999999999187. The combination of multiple algorithms to optimize the output was used for controlling IWRC by Arduino. Thus, the proposed system can improve the efficiency and accuracy of sustainable agriculture.

Keywords

• Artificial intelligence of things (AIoT)
• multiple linear regression (MLR)
• particle swarm optimization (PSO)
• adaptive neuro-fuzzy inference system (ANFIS)
• embedded fuzzy logic library (eFLL) smart farming