Logistic Regression and Internet of Things Based Smart Irrigation to Predict Crops Water Need

Abstract

Irrigation is crucial to agriculture. Due to advancements in technology, when we go outside or whenever crops need to be watered, we no longer need to rely on someone to perform irrigation. Many researchers have attempted to irrigate crops automatically, but accuracy, timeliness, and cost concerns are rarely addressed and given top priority. The proposed approach carries out autonomous watering, which results in smart irrigation, using a wireless sensor network, real-time sensors, and irrigation system control. By using this method, waste is decreased and the necessary water is kept in the container. Instead than just evaluating soil moisture, automated irrigation takes into account the type of crop, the weather, and soil moisture to determine whether the crops need to be watered. By taking into account the aforementioned three factors, a machine learning technique called logistic regression is utilized to estimate the need for water. The logistic regression model, which is based on the values of the three parameters given, forecasts the watering needs of the plants using an arduino-based IoT circuitry. The strategy outlined is advantageous in terms of accuracy, timeliness, and cost. The results of the proposed model has proven its betterment in performance and the amount of automation over the existing irrigation systems. The IoT and machine learning combined model is useful from the point of view of its accuracy, cost and timeliness in predicting water needs for crops as well as full automation is enabled in the irrigation system with this approach.