Computation offloading game for an UAV network in mobile edge computing
Résumé
Due to the limitations of mobile devices in terms of processing power and battery lifetime, cloud based solutions offer an attractive approach to answer these shortcomings. Since offloading intensive computation tasks to an edge/cloud server would achieve impressive performances, computation offloading paradigm has attracted the focus of many research groups in the last few years. This paper considers the problem of computation offloading while achieving a tradeoff between execution time and energy consumption. The proposed solution is intended for a fleet of small drones that are required to achieve highly intensive computation tasks. Drones need to detect, identify and classify objects or situations. Thus, they are brought to deal with intensive tasks such as pattern recognition and video preprocessing. The latter implement very complex calculations and typically require dedicated and powerful processors, which would definitely accentuate the dilemma between energy and delay. We adopted a game theory model where the players are all the drones in the network with three possible strategies. We defined the cost function to be minimized as a combination of energy overhead and delay. The simulation results are very promising and the achieved performances outperformed their counterparts in terms of average system wide cost and scalability.
Mots clés
Mobile communication
Wireless communication
autonomous aerial vehicles
cloud computing
computerised control
energy consumption
mobile computing
mobility management (mobile radio)
synchronisation
telecommunication power management
energy overhead
game theory model
video preprocessing
pattern recognition
edge-cloud server
offloading intensive computation
cloud based solutions
battery lifetime
processing power
mobile devices
UAV network
unmanned aerial vehicles
computation offloading game
Computation Offloading
Mobile Edge Computing
Game Theory
UAV (Unmanned Areal Vehicles)