Metaverse-Based Precision Agriculture: Integrating IoT, AI, and Data Analytics for Sustainable Development
TL;DR
Based on the experimental results, it is concluded that the hybrid CNN-LSTM model is superior and the RMSE, MAE, MAPE and R2 values are 0.829, making it a valuable model for sustainable and accurate irrigation management.
Abstract
The fast development of digital technologies has changed the contemporary agriculture, making it more precise, data-driven, and sustainable in managing the farm. Nevertheless, farmers continue to experience difficulties in terms of observing the process of soil dynamics, optimization of irrigation, and real-time visualization of the state of the fields, which results in the inefficiency of resources and the instability of yields. In order to overcome these constraints, the study suggests the implementation of a combined metaverse-based precision agriculture system comprising of IoT sensing, artificial intelligence-enhanced prediction, and advanced data analytics. A 6-in-1 agro sensor has been used to collect real-time soil parameters such as pH, EC, moisture, temperature, phosphorus, and potassium and utilize them in a Raspberry Pi IoT platform. The Karnataka soil moisture dataset is then trained through CNN, LSTM and CNN-LSTM models after cleaning, smoothing and normalization to classify and predict soil moisture. The system is represented in an interactive metaverse system as an immersive system to monitor and make decisions. Based on the experimental results, it is concluded that the hybrid CNN-LSTM model is superior and the RMSE, MAE, MAPE and R2 values are 0.142, 0.810, 27.342 and 0.829 respectively, making it a valuable model for sustainable and accurate irrigation management.