Amir Hussin, Amir Aatieff and Zainuddin, Ahmad Anwar and Nik Mohd Kamal, Nik Nor Muhammad Saifudin and Mohammad Noor, Normawaty and Mohd Razali, Roziawati and Ayub, Mohd Nor Azman and Harman, Muhammad Farouk and Hassan, Mohd Khairul Azmi and Mohd Tamrin, Mohd Izzuddin and Subramaniam, Krishnan and Kamarudin, Saidatul Izyanie (2026) Towards intelligent aquatic health monitoring in Malaysia's coastal waters: AI-driven harmful algal bloom forecasting with IoT and cloud infrastructure. Open International Journal of Informatics (OIJI), 14 (1). pp. 165-176. E-ISSN 2289-2370
![[img]](https://irep.iium.edu.my/style/images/fileicons/application_pdf.png) |
PDF
- Published Version
Restricted to Registered users only Download (596kB) | Request a copy |
Abstract
Harmful Algal Blooms (HABs) pose significant threats to aquaculture, marine ecosystems, and coastal economies, requiring timely and reliable monitoring approaches for early detection and response. Conventional water quality monitoring methods are often limited by high operational cost, delayed data acquisition, and insufficient forecasting capability for dynamic coastal environments. This study presents an integrated intelligent aquatic health monitoring system that combines Internet of Things (IoT)-based water quality sensing, cloud-based real-time data management, and artificial intelligence (AI) models for HAB monitoring in Malaysian coastal waters. The proposed system employs multi-parameter sensors connected through ESP32 microcontrollers for continuous monitoring of key water quality indicators, with data transmission via MQTT to a cloud dashboard for visualisation and remote access. Field validation was conducted at Sungai Geting, Kelantan, by comparing prototype sensor readings against benchmark YSI ProDSS measurements. Two AI models, namely an Adjusted Combined Model (ACM) integrating Radial Basis Function Networks (RBFN) and Fuzzy C-Means clustering, and Long Short-Term Memory (LSTM), were evaluated for chlorophyll-a forecasting and HAB prediction. Experimental results showed that ACM achieved superior short-term predictive performance with lower RMSE and MAE, while LSTM demonstrated competitive performance for temporal sequence modelling. The findings demonstrate the potential of integrating IoT and AI to support cost-effective, real-time, and predictive HAB monitoring for sustainable aquaculture management and coastal environmental surveillance.
Actions (login required)
 |
View Item |