Artificial Intelligence Techniques for Environmental Weather Monitoring and Prediction System Using IoT and Multi-Model Progressive Dense Self-Attention: Trends and Challenges

Abstract

Environmental weather monitoring and prediction have become increasingly critical due to climate change, extreme weather events, and the need for real-time decision-making across sectors such as agriculture, disaster management, and smart cities. Traditional numerical weather prediction (NWP) models often struggle with high computational complexity and limited adaptability to localized conditions. The integration of Artificial Intelligence (AI) and the Internet of Things (IoT) has significantly enhanced the accuracy, scalability, and responsiveness of weather monitoring systems. IoT devices equipped with sensors collect real-time environmental data, including temperature, humidity, pressure, and wind speed, while AI models such as deep learning, recurrent neural networks, and attention-based architectures analyse complex temporal and spatial patterns. Recent advancements in multi-model progressive dense self-attention networks have further improved prediction accuracy by enabling adaptive feature extraction and long-range dependency modelling. These models effectively handle heterogeneous and high-dimensional data collected from distributed IoT networks. However, challenges such as data quality, scalability, energy efficiency, and model interpretability remain significant barriers. This paper presents a comprehensive review of AI-driven IoT-based weather monitoring and prediction systems, focusing on trends, architectures, and challenges between 2020–2023. A detailed comparative analysis is conducted to evaluate different techniques, followed by insights into future research directions.