Irawan Dwi Wahyono, Aripriharta, Gwo Jia Jong, Khoirudin Asfani, A.N. Afandi, Irham Fadlika
The development of the use of wireless sensor network technology is developing very rapidly, especially in the fields of monitoring, tracking, disasters, and others. One of the uses of WSN is monitoring potential flood disasters. Information provided by wireless sensors is very useful in disaster early warning and management in dealing with post-disaster. The distribution of wireless sensors that are not evenly distributed in flood-prone areas requires grouping sensors based on specific locations so that the information obtained is fast and accurate. This study develops the method of grouping wireless sensors and predicts the potential for flood disasters in river areas. This method uses a combination of the k-mean algorithm and decision tree. The k-mean method is used to clustering the position of wireless sensors in the upstream, middle and downstream of the river flow. After clustering the position of the wireless sensor. Data on each wireless sensor grouping is taken to predict potential flooding in each upstream, middle and downstream area of the river. The results of this study obtained the position accuracy of wireless sensors which obtained an accuracy of 94%. While the average error in predicting flood potential is 3.3%. © 2019 IEEE.
Electrical Engineering, Universitas Negeri Malang, Indonesia; National Kaohsiung University of Science and Technology, Taiwan
