Fuzzy Logic-Based Flood Detection System Using Lora Technology

Floods can occur in places where precipitation and river runoffs are high. Prolonged rainfall over a region is the primary cause of floods. However, there are other factors that influence the severity of the flood such as the geographical location, the soil infiltration capacity and the level of water runoff across a localized area. This makes flood detection systems a very complex system to design as it deals with many variables and ambiguities. Fuzzy logic systems are known to be able to deal with ambiguous systems. The underlying principles of fuzzy logic are, that it is able to imitate the human reasoning process through the combination of linguistic terms and degree of membership. Furthermore, through this approach, the development time and complexity of flood detection system design can be further reduced. In this study, the inputs for this Fuzzy-Logic Based Flood Detection System were rainfall amount, soil moisture, surface water runoff level, and the type of monsoon season Whereas, the output was flood risk. Semtech SX1278 LoRa chip was used to provide a low-power and longrange wireless communication method for the offsite sensor nodes (Branch Nodes). Overall, the system had proven to be reliable as it produces different flood risk levels within expectations when the branch nodes were placed in three distinct elevations. Furthermore, a UI was developed which allows user to view flood parameters using a smart device. Also, the branch nodes are able to achieve more than four months of battery life with a 2000 mAh battery. The flood risk output of this project has novel values as there are currently no publicly available flood risk data in Malaysia that can be compared or categorized at the time this paper was written.