| "Green new infrastructure" is an essential part of the new round of global technological and industrial revolution.It has become an "accelerator" supporting the transformation and upgrading the performance of the Chinese industrial structure.The Internet of Vehicle(LoV),as one of the essential landing application scenarios of the "new infrastructure" project,is a communication network for supporting vehicle and transportation applications.Through massive connections and collaborative computing of the humans,the vehicles,and the road infrastructures,the LoV can enhance the sensing range of vehicles and realize the interconnection of people-vehicle-road-cloud.The IoV deploys a large number of infrastructures with communication capabilities in cities,including the Fifth-generation(5G)base stations,sensors,roadside units,etc.Regarding environmental protection and energy efficiency improvement as the goal of achieving "green".How to build a green vehicle communication network system and how to solve the communication transmission problem under low power consumption are the focuses of this thesis.Firstly,in view of the problem of vehicle-to-vehicle(V2V)communication channel modeling in urban environments,this thesis studies V2V communication channel model based on real geographic location information and real driving environment information.Secondly,in response to the green and high energy efficiency requirements of the IoV,this thesis studies the structural design of nonlinear receivers for IoV communications by introducing low-power radio frequency devices.Thirdly,in view of the capacity problem of the nonlinear Multiple-Input Multiple-Output(MIMO)system in the IoV,this thesis derives the general calculation framework of the uplink achievable rate of the nonlinear MIMO system under the Gaussian channel,and proposes an efficient Antithetic-quasi Monte Carlo algorithm to complete the numerical simulation.Finally,this thesis studies how to perform channel estimation and multiuser detection with nonlinear low-power receivers.In particular,the main contribution of this thesis can be summarized as follows,1.Considering the challenges brought by the modeling of V2V channels in cities,the method of modeling V2V channels based on real geographic location information and vehicle driving environment information is studied.By using communication units deployed at various nodes in the city,the city environment information and vehicle driving information can be obtained and combined with any citys map.Consequently,a real-time V2V channel model based on real geographic location information is proposed.This method adopts real geographic location information to construct a large-scale fading model and employs the vehicles environmental location information to construct a small-scale fading model.After obtaining the two connected vehicles relative moving speed,the Doppler shift can be mo deled.The geometric method is used to identify the types of multiple interactions between vehicles,such as reflection and diffraction.Based on the method,the R-Tree structure is introduced to increase the efficiency of the search.Experimental results show that the proposed method addresses the problem that the channel model does not fit the vehicular networks actual operating environment.2.A structure design of the non-linear receiver for IoV communication was studied to address the problem of excessive high power consumption in the IoV.The non-linear low-power radio frequency(RF)receiver structure was proposed based on the replacement of the intermediate-frequency and high-frequency parts contained in conventional receiver by an amplitude or phase detection receiver.Such alternations solve the defect of excessive power consumption caused by the traditional receivers relying on In-phase/Quadrature(I/Q)modulation.Moreover,we also further analysed the physical characteristics of the proposed low-power non-linear RF receivers,established three mathematical models of non-linear receivers and derived their probability density functions.The approach for deploying low-power non-linear RF receivers on infrastructure in urban environments,such as 5G base stations,roadside units,communication nodes,sensor networks,etc.,to form a green and low-power LoVs ecosystem were also presented in this thesis.3.In view of the challenges imposed by the capacity of non-linear MIMO channels in the Lo V,and considering that the MIMO system composed of non-linear receivers does not have closed solutions,a method for calculating the uplink achievable rate of non-linear MIMO systems is proposed using the mutual information theory.When calculating the uplink achievable rate,the conclusions are drawn from the previous chapter that "due to the presence of reflected signals,the V2V channel can be regarded as a Gaussian channel in dense urban areas with a large number of buildings" is firstly used.Then the uplink achievable rates of three kinds of non-linear MIMO systems are analyzed in the case of Gaussian channels.Numerical simulations are further carried out to verify the analysis.While deriving the general calculation framework,an efficient Antithetic-quasi Monte Carlo algorithm is proposed,which effectively solves the problem of high-dimensional integration.4.In response to the non-linear MIMO receivers challenges in LoV,the Expectation-Maximum(EM)algorithm based on the Gaussian Mixture Model(GMM)is used to solve the channel estimation and multiuser detection problem under the ?-phase observations.Firstly,the problem of channel estimation and multiuser detection in Half Phase Only-MIMO(HPO-MIMO)is modeled as a generalized linear mixture problem under ?-phase observations.Secondly,the GMM algorithm is used to merge multiple Gaussian distributions with different weights to obtain the ?-phase receiver distribution.Then the EM algorithm is used for recovering the signal iteratively.Based on the proposed framework,the GMM-type channel estimator and multiuser detector are designed for HPO-MIMO. |
PDF Full Text Request