AI-Based Energy-Efficient Environmental Monitoring in Precision Agriculture Using LoRa Wireless Sensor Networks

Abstract

The integration of Artificial Intelligence (AI), wireless sensor networks (WSNs), and low-power communication technologies has significantly transformed precision agriculture into a data-driven and sustainable paradigm. In recent years, LoRa-based wireless sensor networks have emerged as a promising solution for large-scale environmental monitoring due to their long-range communication, low energy consumption, and cost-effectiveness. However, challenges such as energy optimization, data reliability, and scalability still persist in real-world agricultural deployments. To address these challenges, advanced deep learning models, particularly optimized Riemannian residual neural networks, have been introduced to enhance predictive accuracy and computational efficiency when dealing with non-Euclidean agricultural data structures. Riemannian residual neural networks extend conventional residual learning frameworks to manifold-based representations, enabling efficient learning from complex, high-dimensional environmental data. These models improve feature extraction, reduce vanishing gradient problems, and optimize energy-aware decision-making processes. Simultaneously, AI-driven optimization techniques enhance network configurations, including packet size, transmission rate, and adaptive communication strategies, thereby significantly reducing energy consumption in LoRa-based systems. This review explores recent advancements in AI-based optimization methods, Riemannian deep learning architectures, and LoRa-enabled environmental monitoring systems. It highlights emerging trends, identifies key technical challenges, and discusses future research directions for achieving sustainable, energy-efficient, and intelligent agricultural monitoring systems.