In an article published in the journal Sensors, researchers from Taiwan and Indonesia developed an innovative system for monitoring and predicting the quality of drinking water using the Internet of Things (IoT) and cloud computing technologies. They aim to reduce the risk of waterborne diseases and ensure the safety of drinking water, which is essential for achieving the Sustainable Development Goals (SDG).
Study: Real-time Water Quality Monitoring and Prediction System using IoT and Cloud Computing. Image credit: Chatchawal Phumkaew/Shutterstock
Background
Drinking water is a necessary resource for human health and well-being, but it can also be a source of various contaminants and pathogens that can cause diseases. According to the World Health Organization (WHO), more than two billion people use a drinking water source contaminated with feces, and 785 million people lack even basic drinking water. Therefore, ensuring drinking water quality is a global challenge and a key target of the SDG.
Drinking water quality depends on several factors, such as water temperature, potential of hydrogen (pH), turbidity, dissolved oxygen, and chemical pollutants. These factors can vary depending on the source, treatment, distribution, and storage of water. Therefore, it is crucial to monitor the water quality parameters regularly and accurately and to alert the authorities and the public in case of any anomalies or hazards. However, conventional water quality monitoring techniques are often costly, time-consuming, and labor-intensive, as they require manual sampling, laboratory analysis, and data processing. Moreover, they cannot provide real-time and continuous information, which limits the ability to respond quickly and effectively to water quality issues.
About the Research
In the present paper, the authors designed a smart water quality monitoring and prediction system based on cloud computing and IoT technologies. Their device consists of four main components: sensors, microcontrollers, web servers, and machine learning models, and can collect, transmit, store, and analyze water quality data in real-time.
IoT is a network of physical devices, sensors, and actuators that can communicate and exchange data over the internet. It has many applications in various domains, such as smart homes, smart cities, smart agriculture, and smart health. One of the potential benefits of IoT is its ability to improve the management and efficiency of water resources by enabling remote monitoring, control, and optimization of water systems. Cloud computing is a paradigm that provides on-demand access to shared computing resources, such as servers, storage, networks, and software, over the Internet. It can offer advantages such as scalability, flexibility, reliability, and cost-effectiveness for IoT applications by providing data storage, processing, and analysis capabilities.
The sensors were deployed in the water sources to measure the water quality parameters, such as temperature, pH, turbidity, and conductivity. An Arduino board-based microcontroller collected the sensor data and sent them to the web server via the Narrow Band-IoT (NB-IoT) network, a low-power and wide-area wireless communication technology. The web server, hosted on the cloud, stores the data in a database and displays it on a web-based dashboard created in an open-source data visualization and monitoring platform, Grafana.
The presented tool also incorporated machine learning models to predict whether the water is drinkable or not based on the collected sensor data. The study compared five different machine learning algorithms: decision tree (DT), gradient boosting (GB), random forest (RF), neural network (NN), and support vector machine (SVM). Additionally, the researchers used a dataset of 10,000 records obtained from the sensors and split it into 60% for training, 20% for validation, and 10% for testing. Moreover, they evaluated the performance of the models using several parameters such as accuracy, precision, recall, and F1-score metrics.
Research Findings
The outcomes showed that the DT model achieved the best performance among the five models, with an accuracy of 99.8%, a precision of 99.8%, a recall of 99.8%, and an F1-score of 99.8%. The GB, RF, NN, and SVM models had slightly lower performance but still achieved a high accuracy of over 98%. The authors attributed the superior performance of the DT model to its simplicity, interpretability, and robustness to noise and outliers.
The developed device could provide real-time notifications to users via messaging applications like WhatsApp or Line in case of unsafe water conditions. The users could also access the web dashboard to view the historical and current water quality data, as well as the prediction results.
Furthermore, the proposed system could be applied to various scenarios and domains, such as household, industrial, agricultural, and environmental water quality monitoring and management. It could help the authorities and the public to ensure the safety and quality of drinking water, prevent and control waterborne diseases, and optimize the water treatment and distribution processes. Moreover, it could also contribute to achieving the SDG, especially goal 6 of ensuring the availability and sustainable management of water and sanitation for all.
Conclusion
In summary, the paper demonstrated the feasibility and effectiveness of using IoT and cloud computing technologies to monitor and predict water quality in real-time. The novel system can provide accurate, reliable, and timely information and alerts to the users, and support the decision-making and actions for improving the water quality and health outcomes. Moreover, it can be extended and integrated with other smart technologies, such as blockchain, big data analytics, and edge computing, to enhance the security, scalability, and efficiency of the water quality monitoring and prediction system.
The researchers acknowledged limitations and challenges in their research, including the dataset size and the limited number of machine learning algorithms explored. They suggested that future work should investigate additional algorithms and expand the dataset for better generalization.
Enhancing Maritime CV with Domain Knowledge