Cooperative Caching Schemes And Coding Aided Data Transmission In Vehicular Networks

With the development of wireless communication technology,the mobile data traffic is drastically increasing in recent years.In addition,around 70% of the mobile data comes from mobile video and this percentage is also increasing.As a typical application scenario,the data volume to be transmitted,especially video data in the vehicular network is also increasing year by year.How to efficiently transmit so massive amount of data in vehicular networks brings challenge and opportunity for the improvement of wireless networks.As a very promising technique to mitigate the transmission burden of the network as well as improve the Quality of Experience(Qo E)of retrieving contents,edge caching receives a lot of attentions in industry and academia these years.In this paper,under the vehicular network,we propose to study and analyze the performance of cooperative caching placement policy and the coding aided data transmission related techniques.Based on the fact that the performance of caching is mainly determined by caching placement policy and the corresponding transmission scheme,our first work proposes a cooperationbased caching policy in the heterogeneous network and characterizes the averaged energy consumption of requesting a content into a function regarding the caching placement policy.The energy consumption function is minimized by optimizing the caching placement design,and analytical results show the proposed cooperation-based scheme outperforms other benchmark schemes when other conditions are fixed.Further,to analyze the impact of transmission scheme on caching performance,in this thesis we apply different transmission schemes under relay based vehicle-to-vehicle(V2V)networks such as automatic repeat request(ARQ)and hybrid ARQ(HARQ)where HARQ includes energy accumulation(EA)and mutual information accumulation(MIA).Specifically,we first model the cache-aided successful transmission probability(CSTP)from nearby V2 V users as a function of caching placement policy for every considered transmission scheme,and then the caching placement policy is optimized to maximize the CSTP performance.Through a series of simulations we can get the optimal caching policy and the corresponding CSTP performance for every transmission scheme.Simulation results indicate that the MIA based caching scheme can reach the best CSTP performance,and the impact of many factors are also investigated.Since most wireless traffic comes from mobile video applications,our second work is to investigate and optimize the video caching in V2 V networks.Specifically,every video is encoded into multiple layers via scalable video coding(SVC)technique.Users can adaptively retrieve one or more layers of data depending on the suffered channel condition and its requirement,and more layers are received,higher video quality the user can enjoy.To further adapt the diverse channel in the transmission,non-orthogonal transmission is applied in which different layers of data are allocated with different powers.Users can decode the maximal layers that fit its channel condition based on successive interference cancellation(SIC)decoding method.Based on the feature of SVC,we first model the enjoyed data rate by the requester when retrieving its requested content from self-caching and V2 V caching which is labeled as cache-aided data rate(CADR).By jointly optimizing the caching placement policy and the corresponding power allocation policy,we can get the maximal CADR value.Simulation results show the non-orthogonal transmission based caching policy can reach higher CADR performance than other schemes.In addition to consider the SVC in the transmission level,we also consider it in the caching placement policy by proposing a layer-based caching scheme in which one caching node may just store part of layers instead of the whole content.Under the layer-based caching scheme,the definition of CADR is re-considered because the caching node may only transmit the cached data layers and the power is allocated within these cached layers.By jointly optimizing the caching capacity allocation policy,caching placement policy,and the corresponding power allocation policy,the maximal CADR performance can be reached.The simulation results show that compared with traditional content-based caching scheme,the considered layered-based caching scheme can get higher CADR performance.In vehicular networks,the transmission of 3D point cloud data is of importance because it is related to the safety of automatic driving.We study the transmission of point cloud data under the vehicle-to-infrastructure(V2I)and vehicle-to-everything(V2X)architectures,respectively.Specifically,the sensed point cloud data is firstly encoded into multiple layers of data,and in the transmission adaptive random network coding(ARNC)technique is applied to adapt the time varying channel conditions.Additionally the scheduling of vehicle-to-vehicle(V2V)and V2 I is studied in the V2 X case.The impacts of vehicle mobility and hard deadline are taken into consideration as well,while the performance of transmission is depicted by total throughput with hard deadline.The throughput performance can be maximized by optimizing the scheduling policy and the design of ARNC packet in each time slot,and simulation results indicate that the ARNC based scheme outperforms other benchmark schemes regarding throughput performance.Finally,this paper analyzes the video update problem.In the vehicular network or in the surveillance network,vehicles or remote monitors need to know the road condition or the monitored scene in real time manner.Thus,a transmitter(base station or local camera)needs to transmit real time video content accordingly.However,due to the uncertain of the channel condition,the receiver may receive nothing or only some lower layers of the video in real manner.This definitely cannot satisfy users’ requirement.For example,a vehicle cannot recognize objects correctly and feedback appropriately from a low quality video.If we want to improve the video quality,then more transmission resource such as time is needed,but the freshness of the sensed video is degraded.Here these exists a tradeoff between video freshness and its quality,that is,higher fresh content may have lower quality while higher quality video may suffer lower freshness.This thesis is proposed to analyze the video update frequency and design the transmission to balance this tradeoff.Specifically,video data is encoded into multiple layers via SVC technique and ARNC is applied in the transmission.We firstly define and model the freshness and quality performance of the video update,both of which are functions of update frequency and transmission scheme.Then a total utility function is established to reflect the impact of freshness performance and quality performance simultaneously.By jointly optimizing the update frequency and the ARNC packet design in each transmission,the utility function can reach its maximal value and the balance between video freshness and quality is achieved accordingly.Simulation results show that compared with other coding aided data transmission scheme,the proposed ARNC based transmission scheme can achieve higher total utility function which means the video freshness and video quality can achieve a better tradeoff.