Disaster-Resilient Emergency Communication With Intelligent Air–Ground Cooperation

Featured by low cost and high mobility, unmanned aerial vehicles (UAVs)-ground assisted communication has been considered as a promising solution to provide fast service recovery for rescue and emergency response with IoT devices across disaster regions. As the terrestrial infrastructures can be annihilated or partially damaged after disaster, the UAV assisted communication system has to be self-organized in highly dynamic and partially observable environment. In this article, we will explore an emergency communication system in post disaster areas with edge nodes and UAV assistance. The system is designed to provide cost effective communication with temporary infrastructures for end users in sophisticated environments. To this end, we first formulate the problem and propose optimal solutions according to observable workloads and communication channel connectivity. Moreover, considering the unknown environment scenario, we further present a cooperative learning-based solution, including hybrid design of edge agents and UAV agents, in which the environmental statistics learned by geo-distributed edge agents can also be utilized by UAV agents for UAV-based node selection and UAV hovering control. Through extensive experiments, we demonstrate the superiority of our proposed solutions in terms of service quality and energy conservation in diverse environments.