Research On Video Semantics-driven Resource Allocation Algorithms In In The Internet Of Vehicles

With the explosive growth of video transmission services in the process of smart networking of vehicles,How to reasonably allocate transmission resources and improving the reliability and accuracy of the output results will have a direct impact on the development of the Internet of vehicles,which is a problem that needs to be solved.The The traditional resource allocation alogorithm is based on Quality of Service(QoS)or Quality of Experience(QoE)transmission resource allocation.These traditional resource allocation methods do not focus on the specific content of the video and cannot meet the needs of video content understanding for services such as autonomous driving in the I-NET environment.Therefore,it is necessary to study the resource allocation optimisation criterion for video content understanding,as well as the corresponding optimisation models and algorithms,to reduce the pressure in the resource allocation process,and to make the needs of processing tasks are effectively met.This study mainly discusses how to allocate resources based on solavage analysis of video content.The research of the thesis is focused on the research of resource allocation in Telematics,the contents of this study include.1.The thesis summarizes the current state of research on resource allocation methods in vehicular network.Firstly,it points out that the phenomenon of massive video data needing to be analysed in the Telematics scenario is elaborated,and on this premise,the development of Telematics is reviewed,including the architecture of Telematics and some key technologies,then the research status of communication resources,computational resources and joint resource allocation in Telematics is introduced,the existing problems and The specific direction of the study is illustrated.Next,the development status of how to use machine learning technology in the field of allocating resources is summarized,which provides support for the rear research understanding-based resource allocation algorithms in the Telematics scenario proposed in this paper.2.In view of the limitation of traditional resource allocation methods,a communication new algorithmic model for the content-based allocation of resources is developed.Firstly,taking the target detection task as an example,the video content understanding accuracy model under the data rate constraint is designed,and then an optimization problem of joint video content understanding accuracy and energy consumption under the communication resource constraint is established.Then,a deep Qnetwork-based resource allocation algorithm is proposed to solve such non-linear problems,considering the real-time nature and environmental changes in the connected vehicle scenario.Finally,the results of communication resource allocation based on video content understanding in the Telematics scenario are analysed through MATLAB simulations,and the performance of the proposed algorithm is evaluated through algorithm comparison.3.In response to the need of content analysis not only for communication resources but computational resources and cache resources et caetera.,a multi-dimensional resource allocation based on video content understanding is proposed it is hoped that the video data can be effectively understood and improve the accuracy of the analysis results the latency requirements.Firstly,the differences between joint allocation and communication resource allocation are analysed,and the impact model of computational and cache resources on detection accuracy is analysed,and the multidimensional resource allocation algorithm based on video content understanding is proposed.Finally,the results of communication resource allocation based on video content understanding in Telematics comparative analysis based on simulation tools illustrates the effectiveness of the proposed algorithm.This study proposes a new algorithm which allocates communication resources based on video contentenvironment in the study of video content understanding-based resource allocation algorithms in the context of resource constraints it can effectively improve the understanding ability of the algorithm,and can also control the time delay.