Search for a command to run...
Efficient irrigation scheduling is essential for sustainable agriculture, particularly in regions facing water scarcity and climate variability. Environmental Sensor Networks (ESNs), integrated with Internet of Things (IoT) technologies, enable continuous monitoring of soil moisture, temperature, humidity, rainfall, and other agro-meteorological parameters. The large volume and temporal complexity of such sensor data have motivated the adoption of deep learning techniques for intelligent irrigation decision-making. This paper presents a critical survey of deep learning-driven irrigation scheduling systems that utilize environmental sensor networks. It reviews commonly used architectures such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, Gated Recurrent Units (GRUs), and Deep Reinforcement Learning models, highlighting their applications in soil moisture forecasting, evapotranspiration estimation, crop water stress detection, and automated irrigation control. The survey compares cloud-based and edge-based deployment models and evaluates performance metrics, datasets, and feature engineering strategies used in recent studies. Furthermore, this paper identifies key challenges including data heterogeneity, sensor noise, limited labeled datasets, energy constraints, and model generalization across diverse agricultural conditions. Finally, potential research directions such as explainable AI, transfer learning, multimodal data fusion, and adaptive reinforcement learning for autonomous irrigation optimization are discussed. This survey provides a comprehensive foundation for researchers and practitioners aiming to design intelligent, data-driven irrigation management systems.