IoT-Driven Agricultural Intelligence: Optimized Double Integral Enhanced Zeroing Neural Network Based Soil Monitoring and Crop Prediction System

Agricultural productivity is increasingly affected by climate variability, soil nutrient imbalance, and inconsistent crop management, often leading to inaccurate crop yield predictions. Addressing these challenges requires real-time monitoring and intelligent prediction systems capable of handling dynamic environmental and soil conditions. In this manuscript, IoT-Driven Agricultural Intelligence: Optimized Double Integral Enhanced Zeroing Neural Network based Soil Monitoring and Crop Prediction System (IoT-DAI-DIEZNN-SMCP) is proposed to address inconsistencies in crop yield prediction caused by environmental variations and soil nutrient imbalance. The system collects real-time sensor data, including soil nutrients such as phosphorus, potassium, nitrogen, and pH, along with environmental factors such as humidity, rainfall, and temperature, which are transmitted to a cloud platform for analysis. The data are pre-processed using Domain Transform Filtering to remove noise and fill missing entries, and then analyzed using the Double Integral Enhanced Zeroing Neural Network. To optimize network parameters, the Reptile Search Optimization Algorithm fine-tunes the model for improved performance. Experimental outcomes show that the IoT-DAI-DIEZNN-SMCP achieves 93.3% accuracy, compared to 90.2% for ACY-MDF-MLAC, 92.3% for IoT-CPP-RCC, and 91.1% for DL-IoT-CPC, demonstrating a clear improvement in predictive capability and reliability for real-time precision agriculture applications.