IoT-Based Smart Farming for Sustainable Agricultural Productivity

Abstract

To meet the needs of a human population exceeding 9 billion by 2050—while facing the challenges of a diminishing freshwater resource, eroding arable land and greater climate variability—the total production of the agriculture sector worldwide needs to increase by some 60- 70% by this year. The Internet of Things (IoT)-based smart farming has emerged as one of the most prominent approaches to reconciling these conflicting forces and enabling real-time field monitoring, closed-loop irrigation, fertigation, and data-enabled crop-stress response. A systematic, PRISMA guided review of 99 studies on IoT based precision and smart farming systems is presented and a four-layer IoT system is proposed and evaluated, which includes perception of soil, weather and canopy sensors, network of LoRaWAN/ NB-IoT connectivity, Cloud management of LSTM soilmoisture forecasting and Randomforest irrigation recommendation, and application and actuation of automated precision drip and fertigation (variable-rate) control and deployed and evaluated in a 30 day field monitoring window over drip irrigated vegetable plot. This water saving agrees with the range of 20-35% reduction in irrigation water use and 3-15% increase in marketable crop yield that have been reported from the various studies looked up that have been conducted in these types of fields and implemented the proposed system. The mochimedi LSTM baseline soil-moisture forecasting model with a 24-hour look-ahead time horizon showed an RMSE of 1.6% VWC and R2 = 0.96, outperforming baselines in linear regression, decision-tree, support-vector-regression, and Random-Forest. These findings show that a low-cost, low-power sensor-to-actuation architecture can deliver measurable, reproducible gains in water-use efficiency and productivity, making it a viable, scalable, and implementable option for IoT-based smart farming to support sustainable agricultural intensification. This paper discusses deployment barriers related to connectivity infrastructure, sensor calibration and maintenance, data quality, and affordability for smallholders, and suggests directions for future research.