Paper Details

Name: Aman Raj, Himanshu Kumar, Aman Prakash Sharma & Annu Priya
Country: India

Year: 2026
Volume: Volume No: 13, January, Year: 2026 (Special Issue)
Page Number: 895-905
DOI: https://doi.org/10.5281/zenodo.19144370

ABSTRACT—
A time-predictive smart irrigation analytics platform that employs Internet of Things (IoT) sensor networks and an adaptive reinforcement learning algorithm (ARLA) for making real-time decisions. The platform combines IoT sensors, deep Q-networks (DQN), and an adaptive exploration method via an epsilon decay function to facilitate improved environmental adaptation feedback, for example, soil moisture and plant health. The system uses multi-agent coordination to allow multiple agribots to learn in parallel to optimize irrigation techniques. The model was deployed and verified to make irrigation schedules predictive using sensor values, providing automated decision making and remote monitoring. By managing sensor data—temperature, humidity, and water level—with actuator control of fans and water pumps efficiently, the system allows highly efficient, data-based irrigation recommendations. By optimizing precision agriculture, the advanced technique raises cost-effectiveness, sustainability, and efficiency. The study also highlights the importance of ethical AI and human-AI cooperation in agricultural automation, outlining key trends, obstacles, and potential future developments.

Keywords—Machine Learning, Precision Farming, Water Management, Adaptive Reinforcement Learning (ARLA), Multi-Agent Coordination.