Optimizing UAV task processing in disaster response with lyapunov-based edge computing

Rakan Armoush

School of Computing and Engineering

Supervisors:

Dr Shidrokh Goudarzi

School of Computing and Engineering

Dr Nasser Matoorianpour

School of Computing and Engineering

Dr Alireza Esfahani

School of Computing and Engineering

This paper introduces a novel cooperative task assignment model tailored for unmanned aerial vehicles (UAVs) leveraging edge computing, aimed at enhancing data collection and processing efficiencies during disaster scenarios. The primary objective is to optimise end-to-end delay and energy consumption across UAV trajectories, critical factors in emergency response operations. I present a Lyapunov-based model that strategically addresses these optimisation challenges. A pivotal contribution of this work is the development of a Lyapunov Optimisation based task processing for queue-based offloading algorithm that efficiently manages task computations by dynamically balancing between local processing and edge offloading based on real-time assessments of system states and resource constraints. Simulations validate the effectiveness of the proposed algorithm, demonstrating notable improvements over existing methodologies in terms of reducing the overall cost of operations. These enhancements are crucial for deploying UAVs in disaster management, providing a robust framework that ensures rapid, reliable, and energy-efficient data handling capabilities.