An advanced automated system using AI and Mesh networks for olericulture farming based on hydroponic method
|
Full Text |
Pdf
|
|
Author |
Adam Wong Yoon Khang, Ahmad Idil Bin Abdul Rahman, Johar Akbar Bin Mohamat Gani, Navin Raj A/L Saravanan, Jamil Abedalrahim Jamil Alsayaydeh, Johar Akbar Bin Mohamat Gani, Albert Feisal Muhd Feisal Bin Ismail and Jaysuman Bin Pusppanathan
|
|
e-ISSN |
1819-6608 |
|
On Pages
|
317-324
|
|
Volume No. |
21
|
|
Issue No. |
5
|
|
Issue Date |
May 10, 2026
|
|
DOI |
https://doi.org/10.59018/032640
|
|
Keywords |
AI, Mesh, IoT, MATLAB, hydroponic.
|
Abstract
In today's globalized landscape, technological strides have revolutionized olericulture farming, prompting a shift from traditional methods to innovative solutions. Hydroponics allows for soil-free crop cultivation but often faces challenges like high labour demands, limited scalability, and inefficient resource management, which hinder its broader adoption. The novelty of this study lies in the development of an advanced automated hydroponic system that uses Artificial Intelligence (AI) and mesh networks to support olericulture farming operations. The goal of this research is to enhance the monitoring and control abilities with prediction functionality features. The method involves ESP32 nodes connecting to the mesh network and subsequently to the Raspberry Pi 4 via Wi-Fi, enabling wireless communication. Utilizing the Message Queuing Telemetry Transport (MQTT) protocol, the nodes transmit water level and crop height data, ensuring redundancy in case of node failure. The Raspberry Pi 4 monitors water levels and triggers alerts through the Blynk application, activating the water pump to maintain optimal conditions. Sensor data is stored in a CSV file, compiled using Python, and imported into MATLAB to generate a predictive model using Neural Network algorithms in an offline manner. Results showed a 25% reduction in water usage compared to traditional hydroponic control systems without predictive modeling and mesh network, hence demonstrating superior efficiency and accuracy. Results also reveal from the prediction model, including RMSE: 0.51387, R-squared: 0.35, MSE: 0.26407, and MAE: 0.34951, which indicate improved predictive accuracy. This study demonstrates the integration of both AI and mesh networks into hydroponic systems, which can significantly improve their reliability and further contribute to the advancement of efficient and resilient hydroponic systems. Future work will refine AI algorithms and expand the system's capabilities in a real-time manner for a wider variety of crops and farming conditions.
Back