PERFORMANCE ANALYSIS OF LORAWAN IN AN INTERNET OF THINGS-BASED FLOOD MONITORING AND WARNING SYSTEM

: Master of Science by Research

Abstract

Floods are unavoidable phenomena that can cause massive loss of people's lives and the destruction of infrastructure. Flash floods rise rapidly in a flood-prone area, which results in property damage, but the impact on human lives is somewhat preventable by the presence of monitoring systems. Although there are many systems widely in practice by disaster management agencies in monitoring flood levels, most of these systems are limited in range. For example, some systems implementing the LoRaWAN have a maximum distance of 300m from the gateway. However, the maximum distance that LoRaWAN can reach is 1.5km. Then, the literature regarding the choice of activation on LoRaWAN for FMWS is limited. Furthermore, in most developing countries, the conventional flood gates in water canals are manually operated and suffer from the lack of real-time monitoring of water levels which might lead to an overflow in the channels and flash floods. On top of that, the lack of real-time data analysis in the system that can be accessed is one of the limitations in Malaysia. Therefore, this research proposes to deploy multiple LoRa-based smart sensors with a LoRaWAN gateway as a network testbed for monitoring flood levels and evaluating the parameter of LoRaWAN. Then, the LoRaWAN’s activation was compared and analysed to identify the best activation for the FMWS. Lastly, the real-time assessment of the risk due to the flood level has been enabled on the Tago.IO dashboard for triggering an early flood warning. The proposed FMWS with LoRaWAN uses an ultrasonic sensor with an Arduino microcontroller to measure water level, LoRa as a communication module, and a single gateway. The end nodes have been tested in several scenarios to test the FMWS’s communication performance in terms of Received Signal Strength Indication (RSSI), Signal Noise Ratio (SNR), delay, and the Percentages of Data Received (PDR). The design of the sensing node involved the hardware and software with the solar panel as the power source. A 3D model for the end node was developed for casing the sensing node. The testing area for testing the performance of LoRaWAN is a 2km radius. Throughout the testing, the proposed system communicates up to 2km in single and multiple node cases. But the data loss is greater than 50%, making it unsuitable for FMWS. Then, the multiple nodes exhibit a slightly lower noise tolerance than a single node. However, the performance is different based on the location. Thus, this paper suggests the SF based on the PDR values for each distance. The recommended SF for the first 1km is SF10 and SF 11 for the next 500m based on the performances and power consumption. The recommended activation for FMWS is Activation By Personalization, ABP, since it is over complete control, especially for achieving a high PDR. Lastly, the data on Tago.IO can be accessed via webpages and Tago.IO mobile application. In conclusion, the LoRaWAN have high performance up to 1.5km distance. However, the higher the SF, the higher the network's performance at long distances. The ABP is the activation that is suitable for the proposed FMWS. Lastly, the warning system will trigger once the water level reaches the warning level.

Date of Award	Oct 2022
Original language	English
Awarding Institution	Universiti Malaysia Pahang
Supervisor	Waheb Abdullah (Director of Studies)