Edge AI in LoRa based MESH network

• Main Author: Ng, Xin Hao
• Format: Final Year Project / Dissertation / Thesis
• Published: 2022
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utar.edu.my/4952/1/3E_1806864_FYP_report_%2D_XIN_HAO_NG.pdf; http://eprints.utar.edu.my/4952/

Natural disasters such as floods frequently occur in Malaysia. Internet of Things (IoT)-based flood early warning systems can forecast the cataclysmic flood event and subsequently inform the public to take evacuation action earlier. However, the issue of disseminating critical information remains an open issue if the communication network is broken. This project aims to develop a lightweight Artificial Intelligence (AI) disaster forecasting and a vicinity communication infrastructure, a resilient NerveNet mesh network with Wi-Fi and LoRa. It will disseminate the information about forecasted flood events ahead of time reliably to the designated recipients even if the base station is destroyed due to a flood. Using the NerveNet Hearsay daemon, texts and images can be synchronised wirelessly in multiple NerveNet nodes' databases. Experimental results validate the AI model, network, and database synchronisation performance. The project findings can serve as the guideline for designing an AI flood early warning system in real life.