Aug 2026· International Journal of Science and Research Archive· 0 citations
TL;DR
This survey explores recent innovations in IoT-based Water Quality Monitoring Systems (IoT-WQMS) integrated with Machine Learning (ML) and Deep Learning (DL) to enable real-time, automated water quality assessment.
Abstract
Ensuring access to safe water is vital for public health and environmental sustainability. However, conventional water quality monitoring methods are limited by high costs, lack of scalability, and delayed analysis. This survey explores recent innovations in IoT-based Water Quality Monitoring Systems (IoT-WQMS) integrated with Machine Learning (ML) and Deep Learning (DL) to enable real-time, automated water quality assessment. The review covers architectures utilizing multi-parameter sensors (e.g., pH, turbidity, TDS, DO, temperature), along with essential data processing techniques such as imputation and normalization. Advanced feature selection methods (RF-MOA, ensemble voting) and hyperparameter tuning techniques (QPSO, Grid Search) are discussed for model optimization. ML models like XGBoost, Random Forest, and ANN, as well as DL models such as CNN-LSTM and MS-CAGRU, demonstrate predictive accuracies up to 99.9%, supporting early contamination detection and regulatory compliance. Applications span urban rivers, aquaculture, and groundwater systems, offering actionable insights for efficient and sustainable water management. The paper also addresses key challenges including sensor calibration, data heterogeneity, and model adaptability, highlighting the role of hybrid AI and Explainable AI (XAI) in enhancing system robustness and transparency. This review provides a comprehensive perspective to guide future research and deployment of intelligent water monitoring solutions.
Water quality monitoring is very important in ensuring public safety and sustainability of aquatic ecosystems. However, traditional approaches used in monitoring water quality are not suitable for monitoring purposes due to lack of capability of continuously assessing water quality and rapidly detecting pollution. To solve this problem, this research presents an intelligent IoT-based framework for monitoring the water quality that is integrated with a hybrid deep learning model with 1D Convolutional Neural Networks, BiLSTM, and an attention mechanism. This framework uses parameters such as pH, turbidity, temperature, DO, BOD, and heavy metals for the classification of the degree of pollution of water. In order to overcome class imbalance problem, the SMOTE approach was applied, resulting in a balanced dataset consisting of 2,745 samples. Modern developments in sensing technology, wireless communication, and embedded computing have contributed to an enhanced adoption of IoT applications in the environmental sphere. IoT-based environmental monitoring solutions combine several sensors to gather the physicochemical parameters including pH, turbidity, temperature, dissolved oxygen (DO), biological oxygen demand (BOD), and heavy metals concentration
R. Vishalakshi, Amit Sharma, Gopaluni Yogindra Siva Sai Karthik et al.· 2026 4th International Confe...· 0 citations
An end-to-end Internet of Things framework designed for real-time water quality monitoring and predictive pollution modeling and a hybrid machine learning architecture—combining Long Short-Term Memory (LSTM) networks for time-series forecasting and Random Forest models for anomaly classification—is proposed.
Parvathy Krishna V, G. S, Sahala Mehrin et al.· International Journal of Tec...· 0 citations
With the worsening environment quality credited to rapid urbanization and industrialization, ensuring optimal indoor air quality is a critical challenge with far-reaching implications for public health, environmental sustainability, and overall wellbeing. Despite its importance, real-time air quality monitoring solutions (AQMS) remain underdeveloped, leaving significant gaps in the ability to analyze and predict environment quality effectively. In this paper, we have presented the development of a Machine Learning Internet of Things (ML-IoT) enabled AQMS equipped with diverse sensors capable of sensing vital environmental parameters which include carbon monoxide, carbon dioxide, particulate matter PM2.5, PM10, temperature, and humidity. We deployed the developed AQMS in an indoor environment and collected environmental data from sensors. A comprehensive preprocessing pipeline, including outlier removal using the interquartile range method and feature scaling, was applied to improve data quality. Three machine learning models including Linear Regression (LR), Random Forest (RF), and Bidirectional Long Short-Term Memory (BiLSTM) were implemented for predictive analysis. Using state-of-the-art data analytics, we have tested these ML models to uncover trends, identify correlations, and predict air quality metrics with improved accuracy. To ensure methodological rigor, 5-fold cross-validation was applied to LR and RF models, while time-series cross-validation was used for BiLSTM to preserve temporal dependencies. The results show that LR achieves high accuracy for temperature prediction with
R
2
of 96.45% and cross validation
R
2
of 95.4%, while RF provides moderate performance for CO
2
prediction with
R
2
of 52.7% and 64.05% in cross validation. BiLSTM improves CO
2
prediction with
R
2
of 89.9%) under standard evaluation and achieves an average
R
2
of 85.4% under time-series validation, demonstrating generalization. To provide real-time visualization of air quality parameters, we have designed a user-friendly dashboard, allowing stakeholders to monitor real-time conditions and derive actionable insights. The developed AQMS finds promising application in industries, houses, office and residential buildings, allowing predictive environment quality monitoring and triggering alarm in case of any anomaly.
Tanzila, Sundus Ali, M. Aslam et al.· Frontiers in Environmental S...· 0 citations
This systematic review analyzes 52 studies (2018–2026) on integrated
IoT and Machine Learning systems for water quality prediction in
aquaculture. Results show that Random Forest achieves accuracy above
94 % in survival classification, while hybrid Deep Learning models such
as CNN-SA-BiSRU reach R² = 0.9765 for dissolved oxygen prediction.
The most monitored parameters are pH (98.2 %), temperature (92.9 %),
and dissolved oxygen (62.5 %). Critical gaps are identified: scarcity of
species-specific datasets for larvae and fry, predominance of single
target models, and limited validation under real production conditions.
The review concludes that IoT-ML integration has demonstrated
technical feasibility, but further applied research is required for effective
adoption in commercial aquaculture farms, particularly in resource
constrained rural settings where most small-scale fish farming occurs.
Jorge Cein Villanueva-Guzmán, Ezequiel Gómez-Domínguez, Julio Cesar Romellón-Cerino et al.· Journal of Computational Sys...· 0 citations
Aquatic environments require continuous monitoring to ensure ecological balance, species health, and efficient resource management. Existing monitoring approaches encounter significant challenges due to dynamic underwater conditions such as varying illumination, turbidity, and background complexity, leading to reduced visibility and unreliable analysis. To address these limitations, an embedded edge-AI framework integrating EfficientNet-B0 for spatial feature extraction and LSTM for temporal sequence modeling is proposed. The system utilizes the Aquatic Intelligent Monitoring Dataset (AIMD-2026), comprising 12,000 images and temporal sequences representing diverse aquatic conditions. Environmental sensing parameters including temperature, total dissolved solids, pH, turbidity, and water level are incorporated to enhance contextual awareness and improve decision-making. Advanced preprocessing techniques ensure data consistency and robustness under challenging underwater conditions. The proposed framework demonstrates improved detection accuracy, temporal consistency, and reduced misclassification compared to conventional approaches. Evaluation using accuracy, precision, recall, and mean Average Precision confirms superior performance. The integration of visual, temporal, and environmental data enables reliable monitoring of fish species, supports maintenance of water quality, and contributes to sustainable aquatic ecosystem management and biodiversity preservation. The system is computationally efficient, scalable, and suitable for real-time deployment in embedded environments for aquaculture and environmental monitoring applications.
S. Naveen· 2026 7th International Confe...· 0 citations
Reliable and scalable monitoring of inland water bodies remains a critical challenge in regions where laboratory-based measurements are spatially sparse and temporally discontinuous. This study proposes a novel hybrid learning framework that integrates physicochemical observations from the Central Pollution Control Board (CPCB) with multispectral Sentinel-2 satellite features for automated binary water pollution classification across Indian states. Pollution labels are derived from regulatory thresholds on dissolved oxygen (DO) and biochemical oxygen demand (BOD). Unlike prior work that relies solely on random data partitioning, we systematically benchmark classical machine learning and temporal deep learning architectures— including Random Forest (RF), CNN-LSTM, and Transformer-based models—under three validation protocols: random, group-based (state-year), and temporal splits. This deployment-oriented evaluation is a key novelty of the proposed framework. The Transformer model achieves a temporal ROC-AUC of 0.786, while the CNN-LSTM achieves perfect recall under class-weighted training, reflecting a deliberate high-sensitivity configuration rather than overfitting. SHAP-based interpretability analysis identifies spectral bands and turbidity-sensitive indices as dominant predictors. The findings establish a practical and interpretable pathway toward scalable regulatory water quality monitoring under realistic spatial and temporal distribution shifts.
A. K, Keshav Pareek, Atul Raghuvanshi· 2026 7th International Confe...· 0 citations
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
