Wireless sensor networks and machine learning meet climate change prediction

• Published in: International journal of communication systems 2021-02, Vol.34 (3), p.n/a
• Main Authors: Anh Khoa, Tran, Quang Minh, Nguyen, Hai Son, Hoang, Nguyen Dang Khoa, Cao, Ngoc Tan, Dinh, VanDung, Nguyen, Hoang Nam, Nguyen, Ngoc Minh Duc, Dang, Trung Tin, Nguyen
• Format: Article
• Language: English
• Subjects: Agricultural commodities; Agricultural management; Applications programs; Body area networks; Climate change; Cloud cover; Control systems design; Data collection; Data management; Desertification; Internet of Things; linear regression; Machine learning; Natural disasters; Nodes; predict; Regression analysis; Risk levels; sensor; Sensors; Weather stations; web application; Wind speed; Wireless networks; wireless sensor network; Wireless sensor networks
• ISSN: 1074-5351; 1099-1131
• DOI: 10.1002/dac.4687

Summary Climate change is one of the main challenges faced by the development of every country. For countries producing agricultural commodities, the climate affects the quantity and quality of products. Many methods have been proposed to keep track of climate. One traditional method is the weather station model, which indicates the temperature, wind speed, and direction and extent of cloud cover. However, this method of predicting climate change has low accuracy due to geographical variation, for example, mountainous or forested areas. Recently, a combination of wireless sensor networks (WSN) and machine learning (ML) has been considered for prediction with the Internet of Things (IoT), for instance, through a wireless body area network. For climate change prediction, we design and develop a control system that uses node sensors to collect data in sandhills and beaches, with data management conducted via a web application with three components. The first component is designed to collect data from the node sensors. The second component is mainly used to control the system through a web application. The third component uses linear regression in ML to analyze the data to predict weight and volume. The complete system has been tried and tested in real time on a 10‐m2 area of a beach at Binh Thuan province, Vietnam, where sensor node data were wirelessly collected over a cloud using a web application. This enabled assessment of the current state of the land at a coastal sandy beach, as well as prediction of the risk level of desertification and natural disasters. For climate change prediction, we design and develop a control system that combines between wireless sensor networks (WSNs) and machine learning (ML) under the Internet of Things (IoT) environments. The complete system has been tried and tested in real time on a 10‐m2 area of a beach at Binh Thuan province, Vietnam, where sensor node data were wirelessly collected over a cloud using a web application.