IoT-Driven Metaverse Agricultural: A Novel Approach for Precision Farming and Sustainability

Abstract

Due to the current demand for sustainable agriculture and efficient use of resources, more and more modern technologies have been used in agriculture. Issues such as over-watering, over-fertilizing, and sub-optimal yields frequently occur as a result of traditional practices, and are in need of data-driven solutions with exactness. This research introduces an IoT-based metaverse agricultural system for precision farming that leverages real-time monitoring and predictive analytics to enhance the sustainability of agriculture. The sensors are connected to the Arduino Mega microcontroller and the data is transmitted to the cloud through the soil sensors (NPK, pH, moisture, temperature). The data from the sensors is cleaned, smoothed and outliers removed before analysis. For this work, Agriculture IoT 2024 dataset has been selected as it contains different soil types, crop types and environmental parameters which are used for the training and testing of the model. In this paper we propose to use machine learning models, namely MLP, SVM and hybrid ensembles to estimate soil moisture and crop requirements. A hybrid model gives better results with an RMSE value of 0.0229, MAE of 0.0178, and an R-square value of 0.9577. The optimized irrigation and fertilizers led to a 38% reduction in water use and 36% in fertilizer use and to a yield increase of 25%, thus creating a solid, sustainable and effective precision farming methodology.