Development of automated hydroponic system using solar panel and artificial intelligence for agriculture

Kathiresan, Kristisswaran (2025) Development of automated hydroponic system using solar panel and artificial intelligence for agriculture. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Development of automated hydroponic system using solar panel and artificial intelligence for agriculture.pdf - Submitted Version Download (6MB)

Abstract

Hydroponic system is the practice of growing plants using only water, nutrients, and a growing medium without using soil which fosters rapid growth, stronger yields, and superior quality. Addressing challenges such as overlooking optimal growth conditions, failing to crop tracking accuracy in hydroponics, lack of reliability in hydroponic system node redundancy, and wasting energy. The project implements Artificial Intelligence (AI) in tracking and analysing crop growth rates using sensor parameters, optimizing communication within the mesh network, and implementing remote control for precise management. Additionally, solar panels are integrated to save energy and promote sustainability in the hydroponic setup to be specific as an eco-friendly. To enhance data accuracy, particularly in monitoring water levels, the project leverages AI algorithms to process sensor data efficiently. The results will be analysed and optimized for better performance of the system. This project uses a microcontroller, ultrasonic sensor, water level sensor and a solar panel. The system will be integrated with a smart phone so that user can monitor and control easily in real time. The approach involves ESP32 nodes connecting to the mesh network and then to a Raspberry Pi 5 via Wi-Fi for wireless communication. Using the MQTT protocol, the nodes transmit data on water levels and crop height, ensuring redundancy in case of node failures. The Raspberry Pi 5 oversees water levels and sends alerts through a Blynk application, activating the water pump to maintain ideal conditions. Data is recorded in ESP32 nodes will be transferred to the Raspberry 5, where the data will be compiled using Python and imported into MATLAB in itself to process the data for creating predictive model. The Blynk application enables real-time monitoring, providing users with insights into the hydroponic system's status. Result from the prediction model, including RMSE = 0.44701, R-squared = 0.81, MSE = 0.19982, and MAE = 0.166649, indicate improved predictive accuracy. Wide Neural Network and the Narrow Neural Network excelled in handling high-dimensional data but Decision Tree outperformed both with its exceptional predictive accuracy, making it the best fit. This comprehensive approach bridges the gap between identified problems and innovative solutions, contributing to the advancement of efficient and resilient hydroponic systems. With the successful completion of the project, it will establish a sustainable agricultural system that improves food security, environmental conservation, and economic viability in the industry.

Actions (login required)

View Item