An Integrated Network Model with Sequential Approach to Elevate Irrigation in Smart IoT Environment

• Published in: 2025 5th International Conference on Pervasive Computing and Social Networking (ICPCSN) pp. 1941 - 1946
• Main Authors: M, Baskar, Periyasamy, P.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.05.2025
• Subjects: agriculture; Crops; Intelligent sensors; Internet of Things; Irrigation; Long short term memory; moisture; Monitoring; prediction; Predictive models; Rivers; Soil moisture; Water resources
• DOI: 10.1109/ICPCSN65854.2025.11035594

Many countries are endowed with abundant natural resources such as fertile land, rivers, groundwater, fertilizers, and a favorable environment. In numerous regions, agriculture remains the principal source of income for the population. Historically, water resources like groundwater and river systems have been relatively plentiful. However, the primary reason for their inefficient use lies in the inadequate understanding of how to manage these valuable resources effectively. Smart agriculture presents a promising solution by optimizing water application across varying soil types, climatic conditions, and crop growth stages. This is achieved by deploying soil moisture sensors at strategic observation points to monitor water retention levels accurately. Nonetheless, challenges such as sensor malfunctions or communication failures may hinder the acquisition of real-time soil moisture data, thereby affecting irrigation management. The Internet of Things (IoT) offers transformative potential for modern agricultural practices. However, IoT devices often suffer from limited energy capacity and face complex routing challenges. Thus, it becomes imperative to develop a robust, energy-efficient IoT-based irrigation framework for agricultural use. The proposed system incorporates data acquisition and forecasting as its core processes, employing a deep learning-based Sequential Network Model (SNM) that utilizes Long Short-Term Memory (LSTM) units to retain temporal patterns. Initially, benchmark resources are employed to collect essential imagery and sensor data. This data is then fed into the irrigation forecasting module, which supports informed decision-making for irrigation scheduling. Ultimately, this approach empowers farmers to lower production costs while enhancing crop productivity.