Edge AI Based Autonomous UAV for Emergency Network Deployment: A Study Towards Search and Rescue Missions

Abstract

Each year natural disasters claim millions of lives across the globe. The numerous tireless rescue missions are the aftermath of natural disasters such as typhoons, hurricanes, blizzards, forest fires, and heavy storms. Unfortunately, the first responders responsible for rescuing the people in distress get paralyzed in their efforts as the wireless network is the first system to malfunction during such adversities. An intelligent system based on Unmanned Aerial Vehicles (UAV) which helps in locating and communicating with the survivors offers a promising alternative for mission-critical (MC) scenarios. The survivors are located by the autonomous UAV via an edge AI image classifier model. Further, due to the distinctive features such as flexible deployment and rapid reconfiguration, drones can readily change location dynamically to deliver on-demand communications to users on the ground in emergency scenarios. As a result, using UAVs as access point to local area network has been assessed as a practical approach for supplying instantaneous connection in MC situations. The proposed solution here does not require any manual control. It can automatically maneuver, land, and take off using Aruco markers. This work includes a precision landing study and a Received Signal Strength Indicator (RSSI) study of the network provided by the UAV which examines the constraints and applications of the system.