A Study On Relay Selection Algorithms For Multi-hop Device-to-device Communication In Mobile Cellular Networks

With the development of mobile Internet and the emergence of big data,cloud computing and virtual reality technologies,users' demand for data communication has increased dramatically.More and more mobile devices are accessing cellular networks.However,radio spectrum resources are limited.How to provide faster and lower-latency communication services to a larger number of users has become an important issue in mobile communications.In recent years,new services such as video sharing,live games,and proximity services are emerging continuously in cellular networks.To accommodate such new services,3GPP has introduced device to device(D2D)communication in cellular networks in LTE Release 12,and multi-hop D2D communication based on direct D2D communication has drawn a lot of attention in the research community.Relay selection is an important issue in multi-hop communications,which has widely been studied in the areas of mobile ad hoc networks,wireless sensor networks,and computer networks.However,mobile cellular communication has its own characteristics,such as centralized control architecture,diverse user mobility,and different service demands.Therefore,it is an important part to study relay selection in multi-hop D2D communication taking into account the characteristics of mobile cellular communication.This thesis studies the relay selection problem in multi-hop D2D communication in mobile cellular networks.Taking into account user mobility,user activeness,and user content demand in different network scenarios,three different multi-hop D2D communication relay selection algorithms are proposed,namely,mobility and activeness aware relay selection algorithm,content and energy aware relay selection algorithm,and common content request probability and activeness based relay selection algorithm.The thesis first proposes a Mobility and Activeness aware Relay Selection(MARS)algorithm.This algorithm considers the network scenarios where there are a number of candidate relay users and different candidate users have different mobility and communication activeness.In relay selection,it considers the impact of user mobility on multi-hop D2D communication and meanwhile considers user activeness in calculating the effective information transmission rate that a candidate relay user brings to multi-hop D2D communication.The candidate relay user with the largest effective information transmission rate is selected as the relay node.Simulation results show that compared with the traditional random relay selection algorithm,MARS can bring a larger information transmission rate to the system.By comparing the random walk model with the Gaussian Markov model,it is shown that the MARS algorithm is more suitable for the current rate and historical rate relevant mobility model.Then the thesis proposes a Content and Energy aware Relay Selection(CERS)algorithm.This algorithm considers the network scenario where a number of users request for the same content simultaneously.In relay selection,it considers the impacts of the contents that different users request for and the residual energy of a user device on multi-hop D2D communication,and selects the candidate user that requests for the same content and has a larger residual energy as the relay node.Simulation results show that the CERS algorithm can bring a larger effective information transmission rate to the system than the maximum channel gain relay selection algorithm.Moreover,it has a larger advantage in a network scenario where user energy is more relevant,such as in the transmission of a very large file.Finally,the thesis proposes a common content request Probability and Activeness based Relay Selection(PARS)algorithm.This algorithm extends the network scenario where a number of users request for the same content simultaneously to a common multi-hop D2D communication scenario.In relay selection,it considers both the common content request probability and user activeness,and calculates the average effective information transmission rate that a candidate user brings to multi-hop D2D communication.The candidate user with the largest average effective information transmission rate is selected as the relay node.Simulation results show that compared with the CERS algorithm,the PARS algorithm can bring larger information transmission rate and effective information transmission rate to the system.