| Cellular Vehicle-to-Everything(C-V2X)is an important application of the 5th Generation of Wireless Communications(5G)in intelligent transportation field,which is also a vital part of Chinese new infrastructure construction.In C-V2X,vehicle users can acquire the real-time and dynamic connections of "people-vehicle-road-cloud" based on traditional cellular communications and sidelinks.Besides,the high mobility of vehicles further aggregates the uneven temporal-spatial distribution of wireless traffic in C-V2X,leading to the mismatch between radio resources of cellular base station and user requirements.These will lower the energy resource utilization of wireless network and cannot meet the requirement of green communication.With the development of artificial intelligence,machine learning and big data are used to analyse wireless historical data,predict the variation of wireless traffic and user behaviors to improve the utilization of radio resources.Recently,how to use the prior information to design resource pre-allocation strategy has been the key point of handling the mismatch between the user requirement and radio resources in C-V2X.Motivated by these problems,this thesis focuses on resource pre-allocation strategy based on the priori information in C-V2X.First,data analysis is carried out on the vehicle speed distribution from a public dataset and an energy efficient resource pre-allocation strategy is studied.On this basis,the impact of base station traffic volume on energy consumption is further considered,exploiting how to jointly optimize user association,subcarriers and power allocation to minimize energy consumption.The main contributions and innovations of this thesis are summarized as follows:1)An energy efficient resource pre-allocation(EE-RPA)strategy based on priori information in C-V2X is exploited,and the average system energy efficiency(EE)is maximized through time allocation and power allocation with QoS and outage probability constraints.First,data analysis is carried out on the vehicle user velocity from the public dataset to find a proper fitting function.On this basis,to maximize the average system EE,channel distribution information is used to.simplify the outage probability constraints,and a two-layer iterative EE-RPA algorithm is proposed.This algorithm first transforms the fractional form of the objective function into a subtractive form.Then energy efficiency factor is introduced for the outer iteration,and the inner iteration is solved by linear programming and Lagrangian duality theory.Finally,simulations are carried out to verify the performance of the proposed algorithm in different speed distributions of vehicle user equipment(VUE)and different requested data volume.The simulation results show that the kernel density estimation with Gaussian kernel can characterize the speed distribution of VUE.At the same time,the proposed algorithm has good convergence and the proposed algorithm can obtain at least 65%energy efficiency gain compared with the reactive resource allocation strategy.Meanwhile,the unequal number between high mobility VUEs and low mobility VUEs also has an impact on the system performance of the proposed algorithm.2)An energy consumption optimization strategy based on the priori information in C-V2X networks is exploited.By extending the available subcarriers through the frame in prediction window,and utilizing the many-to-one matching theory,a resource pre-allocation strategy based on matching theory is proposed to minimize the system energy consumption by jointly optimizing user’ association,subcarriers and power allocation.Considering the energy consumption model which is linearly proportional to the load of base station,a multi-dimensional resource pre-allocation algorithm based on the matching swap is proposed under the constraints of user quality of service.The proposed algorithm first decomposes the original problem into discrete sub-optimization and continuous sub-optimization problems.Secondly,the available subcarriers of the base station are virtually expanded through each frame within the prediction window and bilateral many-to-one matching model is used to solve user association and subcarrier allocation.Finally,Lagrangian duality theory is used to find the optimal transmit power.Simulation results show that the proposed algorithm has a fast convergence rate and can significantly reduce energy consumption compared with traditional resource allocation strategies. |
