Research On Energy Optimization Of UAV With Mobile Edge Computing

UAV(unmanned aerial vehicles)developed in the 1920s and are often referred to as drones or remotely piloted aircraft.Due to its high maneuverability,ease deployment and low cost,it has been widely used in the past few decades.UAV can be broadly classified into two categories,fixed-wings and rotary-wing.Fixed-wing UAV usually have high-speed and heavy payload,but they must be kept moving at all times(that means they cannot hover in the air),so that they are not suitable for some close-range exploration and search scenarios.In contrast,although rotary-wing UAV has limited moving speed and payload,it can change direction at any time and hover in the air.At present UAV have been widely used in MEC(mobile edge computing),how to reduce the energy consumption during the mission of UAV has become a new challenge.MEC is widely considered as a key technology for realizing the vision of the next-generation of the internet,such as tactile internet(with millisecond-scale reaction time)and the internet of things.At present,researchers from both academia and industry are actively trying to promote the development of MEC technology by pursuing the convergence of technologies and theories in the fields of mobile computing and wireless communications.For large quantities of data that are difficult to calculate in time,the UAV can transmit data to the idle computing device and assist in the calculation to achieve the task of making full use of the idle device and quickly processing the data.This has important theoretical and practical significance for military and civilian applications with particularly high computing power requirements,so it is necessary to carry out work on this aspect of research.Firstly,this dissertation studies the application of data transmission by rotary-wing UAV.The purpose is to minimize the energy consumption of the data transmission via jointly optimizing the UAV's trajectory,speed,acceleration,transmission power and data offload rate under a set of actual environmental constraints.In order to achieve this goal,the models of problem was derived mathematically.Due to there are infinite amount of optimization variables and the some constraints are non-convex,so that it is hard to solve directly by applying convex optimization techniques.Therefore,this dissertation approximate the problem by applying the discrete linear state-space approximation and linearization technology,and proposes an algorithm based on CCCP(concave-convex procedure).The numerical results show that the algorithm can obtain better performance gain under multiple ground stations receive data at the same time.Secondly,this dissertation further combines the fixed-wing UAV with the MEC,and minimize the energy consumption of the whole system via jointly optimizing the fixed-wing UAV's trajectory,data assignment rate as well as CPU's computational speed under the constraints of a series of practical environments.To this end,this dissertation first derive the energy consumption model of data processing and then combine with the energy consumption model of data transmission and fixed-wing UAV's flight.Thus,the models of energy consumption of the whole system was derived mathematically.Due to the objective function contains continuous and coupled optimization variables as well as it is difficult to judge the convexity,so that the problem is hard to solve directly.Therefore,this dissertation firstly discretizes and decouples the optimization problem and then the upper bound of the objective function was derived.Behinds,some non-convex constraints was linearization.Finally,an simple and efficient algorithm was proposed based on the CCCP.Numerical results show that the proposed method achieves great performance gain compared to the original system.