IoT-Enabled Framework for Smart Agriculture: Enhancing Crop Production with Environmental Sensors and Machine Learning
Abstract
The world population is increasingly growing which requires massive improvements in agricultural productivity and sustainability. The common traditional modes of farming have the failure of maximally utilizing resources and a failure to adapt to changing environmental features. This paper presents a new IoT-based system of smart agriculture that is aimed at streamlining production in crops. Such a network is a distributed system that gathers data on soil moisture, temperature, humidity, and pH levels as well as a utility managed with a cloud-based data analytics system. To process the sensor data, we use the algorithm of machine learning, namely, a Random Forest (RF) model and a Long Short-Term Memory (LSTM) network. These models recommend real time irrigation schedules and forecast the possibility of crop stress. Our testbed deployment results indicate that the proposed framework can be used to enhance crop yield and also reduce water use as opposed to conventional methods.