Internet of things for smart farming and frost intelligent control in greenhouses

Abstract This paper is aimed at continuation of frost intelligent control from the smart farming perspective that is proposed by Castaneda and Castano (2017) through Internet of things (IoT) and a Weather Station with an Artificial Neural Network (ANN). Moreover, an intelligent anti-frost irrigation management system is presented. The climatological station and the ecological anti-disaster frost irrigation interact with the environmental system through a website, allowing the real time interconnection, acquisition and monitoring of information through mobile phone systems (GSM/GPRS) and internet (TCP/IP) services. The system is self-sustaining through the use of solar panels. The ANN could be used to optimally predict the inside temperature of greenhouses and a Fuzzy Expert System (FES) controls the activation of a water pump. The ANN input variables involve the relative humidity, temperature of outside air, solar radiation, wind speed and inside air relative humidity. The fuzzy control and ANN allows the prediction of the internal temperature of the greenhouse and the cropland temperature, which are used to activate the anti-frost water distribution system. The ANN models give a temperature prediction through the coefficient of determination of variance analysis (ANOVA) method. The R 2 values for temperature in summer season were 90.23, and 91.30; and for the winter season were 94.28, and 95.22, respectively. The fuzzy associative memory (FAM) controls the activation of the anti-frost irrigation system with five outputs for controlling the climatological frost presence: No-frost (NF), Possible-frost (PF), Mild-frost (MF), Severe-frost (SF) and Hard-frost (HF).