Data And Computing Offloading For Real-Time Situational Awareness Map Application In Vehicular Fog Computing

Recently,with the development and application of 5G,the Internet of Vehicles(IoVs)and autonomous driving have attracted lots of research interests.The concept of vehicle fog computing is also proposed because it can make full use of idle computing and communication resources on vehicles.As a typical application scenario of Internet of vehicles and one of the necessary applications of autonomous-driving vehicles,real-time situational awareness map can help vehicles to achieve accurate positioning and path planning functions.The construction procedure of real-time situational awareness map is mainly composed of two parts:obtaining map data from the cloud and processing of sensing information preceived by local sensors,which respectively need the help of data offloading and computing offloading technologies.However,at present,most of the data offloading studies are not carried out in accordance with the characteristics of map data,resulting in abundant data transmission and excessive energy consumption.And the study of computational unloading seldom really combined the allocation of communication and computing resources jointly when making decisions.In view of the above problems,this thesis focuses on the construction of real-time situational awareness map,and studies the data offloading for cloud map data obtaining and the computational offloading for local sensor data processing,respectively.The main research contents include:(1)For cloud map data acquisition in real-time situational awareness map,this thesis proposes a cluster-based data offloading scheme.The proposed scheme is designed based on the DMOWC(Distributed Multi-Objective Weighted Clustering)algorithm to construct cluster,and adequately considering the characteristics of map data such as closely related to the geographical position,large amount of basic data and frequent update.By sending basic data only once and combining with vehicle’s direction to offload targetly,the proposed data offloading scheme effectively reduce the amount of data transmission and redundancy.Simulation results show that the DMOWC algorithm is suitable for this data unloading scenario,and the proposed data offloading scheme for map data can significantly reduce system energy consumption and data offloading delay.(2)For local sensor data processing in real-time situational awareness map,this thesis proposes a computing offloading decision strategy based on the joint allocation of communication and computing resources.Based on the cluster established by DMOWC,this strategy firstly determines the number of the optimal communication RS(Resource Set)based on the graph coloring method,and then uses the backtracking method with time constraint pruning to solve the optimal computing offloading decision considering the joint allocation of two resources.The simulation results show that the offloading strategy of joint allocation of resources can accomplish more computing tasks offloading by V2V with limited resources and maximize the utilization of resources.