DESIGN OF A LANDSLIDE EARLY WARNING SYSTEM BASED ON THE INTERNET OF THINGS

• Main Author: Pujian Wisesa, Tri
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/76392
• Institution: Institut Teknologi Bandung

Landslides in Indonesia are the natural disaster with the third highest number of occurrences in 2021, hence the need for a landslide mitigation plan. A mitigation plan is needed to minimize losses caused by landslides, namely by developing a landslide early warning system. This research designs an Internet of Things (IoT)- based landslide early warning system to optimize efficiency and accuracy in detecting potential landslides. The system consists of several sensors installed in landslide risk areas to monitor environmental parameters such as rainfall, soil moisture, tilt and soil movement. These sensors will collect data in real-time and send it to a data storage center through an IoT network. At the data storage center, the data from the sensors will be processed using artificial intelligence so as to provide the status of potential landslides. After the data is processed, the warning system will display the data of the measured environmental parameters and the potential landslide status to the monitoring website that has been created. This research produces an IoT-based landslide early warning tool with artificial intelligence that has an error rate of 5.32% in measuring rainfall parameters, 5.26% in soil moisture, 2.9% in soil movement, 1.5% in x-axis tilt (roll) and 1.9% in y-axis tilt (pitch) and with the ability to send data at a maximum distance of 712.55 meters. The results of this research are expected to contribute in efforts to improve mitigation of landslides