• Ayush Verma
• Anish Bammidi
• Shreyash Jitendra Sarade

Efficient water management is a critical challenge in modern agriculture, compounded by climate change and resource scarcity. This paper presents AquaAura, a low-cost framework for predictive soil moisture management using edge artificial intelligence (AI). The system utilizes an ESP32 microcontroller and environmental sensors to execute a lightweight Artificial Neural Network (ANN) on-device, enabling autonomous, low-latency soil moisture prediction without reliance on cloud infrastructure. To ensure model robustness and reproducibility, the ANN was trained on a public, multi-year agro-climatic dataset. A comparative benchmark against five machine learning models—Linear Regression, Random Forest, XGBoost, LSTM, and GRU—demonstrated the proposed ANN’s efficacy, achieving a Mean Absolute Error (MAE) of 3.15% and an R2 of 0.945 on the test set. Post training integer quantization reduced the model size by 74% and inference time by 4x with negligible impact on accuracy, making it suitable for resource constrained microcontrollers. In subsequent field trials, the AquaAura system reduced water consumption by an average of 38.7% compared to conventional timer-based irrigation systems while maintaining optimal soil moisture levels. This research validates that lightweight, on-device deep learning, trained on verifiable datasets, offers a scalable and economically viable solution for enhancing water-use efficiency in precision agriculture.

Precision Agriculture, Edge AI, Soil Moisture Prediction, Internet of Things (IoT), ESP32, TensorFlow Lite, Sustainable Farming, Ar tificial Neural Network (ANN)

"AquaAura: An Edge AI Framework for Predictive Soil Moisture Management Using Public and Field-Validated Datasets", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 9, page no.f30-f39, September-2025, Available :http://www.jetir.org/papers/JETIR2509505.pdf

"AquaAura: An Edge AI Framework for Predictive Soil Moisture Management Using Public and Field-Validated Datasets", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 9, page no. ppf30-f39, September-2025, Available at : http://www.jetir.org/papers/JETIR2509505.pdf