Paper #5

Enhancing Agricultural Sustainability: An IoT-Based RNN-LSTM Model for Precision Sub-Surface Moisture Monitoring and Irrigation Optimisation

Shamala Maniam, Erfan Memar, Tee Yei Kheng, Nisha Kumari, Hin Yong Wong and Mukter Zaman

Abstract: Water directly influences plant growth and vitality and is a critical resource in precision agriculture (PA). Soluble fertilisers are transported to plant roots through irrigation, making precise water management essential for optimising crop productivity and minimising resource wastage. Inadequate or excessive irrigation disrupts nutrient distribution, increases operational costs, and negatively affects crop yield. Accurate monitoring of sub-surface soil moisture, particularly at root depth, is therefore vital for effective irrigation control. This study addresses key limitations in existing PA systems by developing an automated Internet of Things (IoT)-based real-time soil moisture monitoring and irrigation framework integrated with a recurrent neural network (RNN) employing long short-term memory (LSTM) for moisture prediction. Customised sub-surface soil moisture probes equipped with five sensors at different depths were deployed at a real plantation site. The probes utilised time domain reflectometer (TDR) technology to capture high-resolution moisture measurements. Sensor data were transmitted to the cloud using an ESP32-based low-range communication module, forming a wireless sensor network (WSN) across the designated study area. A continuous six-month dataset was collected and analysed to train and validate the proposed RNN-LSTM model. The model demonstrated strong predictive capability, achieving an accuracy of 95 ± 2%, a mean absolute error (MAE) of 0.6362, a root mean square error (RMSE) of 1.1544, and an R² value of 0.3331. These results confirm the model’s effectiveness in capturing sub-surface soil moisture dynamics under real field conditions. Overall, the proposed IoT-enabled predictive irrigation system provides a scalable and data-driven solution for improving irrigation efficiency.

Keywords: Internet of Things; Precision Agriculture; Real-time Monitoring; Recurrent Neural Network.