Authors:
Anishbhai Vahora, Safvan Vahora

Addresses:
Department of Electronics and Communication Engineering, Birla Vishvakarma Mahavidyalaya Engineering College, Anand, Gujarat, India. Department of Information Technology, Government Engineering College, Modasa, Gujarat, India.

Abstract:

This paper outlines a powerful Edge-AI-based IoT platform that uses localised intelligence to transform precision agriculture. Conventional cloud networks typically experience latency and bandwidth limitations in distant farming areas. This framework supports real-time data processing and decision support for irrigation and pest management by running machine learning models at the network edge. The experiment uses a specialised dataset of environmental parameters, including soil moisture, ambient temperature, humidity, and light intensity. To put this architecture into practice, edge computing nodes with Raspberry Pi 4s, sensor arrays, and the TensorFlow Lite library were used to optimise models. The combination of Python-based data processing ensures efficient time and power consumption. Findings have shown that the edge-based architecture offers much lower response times than conventional cloud architectures. The system has been tested using 490 examples of granular field data and shows a prediction accuracy of more than 95 percent for localised irrigation requirements. This study will present a scaled model of smart farming to ensure that insights derived from data are accessible even in regions without internet access, thereby maximising resource use and improving crop production through proactive environmental control.

Keywords: Precision Agriculture; Edge Computing; Artificial Intelligence; Internet of Things; Real-Time Monitoring; Decision-Making Process; Farming Recommendations; Cloud AI System.

Received on: 30/03/2025, Revised on: 08/06/2025, Accepted on: 15/08/2025, Published on: 05/06/2026

DOI: 10.69888/FTSIN.2026.000704

FMDB Transactions on Sustainable Intelligent Networks, 2026 Vol. 3 No. 2, Pages: 65-74

• Views : 15
• Downloads : 5

Download PDF