The Research Of Vr-based Multi-User Cooperative Transmission And Resource Allocation Algorithm In Mobile Edge Network

With the explosive growth of mobile applications,mobile edge networks offer computation offloading services to the user equipment by placing computing and storage resources closer to the network edge.This approach effectively reduces the computational pressure on user devices,reduces latency,and supports the development of more computationallyintensive and latency-sensitive applications such as Virtual Reality(VR).VR provides immersive interactive experiences and holds great potential for various fields,including office,education,and industrial production.To better support VR applications and alleviate the rendering computational pressure of mobile VR devices,high-definition 360-degree panoramic view data with large amounts of data is usually offloaded to mobile edge servers.However,in multi-user VR scenarios,offloading rendering tasks that involve large amounts of view data and have high synchronization requirements to mobile edge servers will result in a significant consumption of computational and communication resources,posing challenges to the mobile edge network.Therefore,the efficient rendering and transmission of VR view data within the constraints of limited computational and communication resources and stringent latency requirements constitute a critical concern that urgently necessitates resolution in VR-based multi-user mobile edge networks.This thesis focuses on the offloading and transmission requirements of view data in multi-user VR scenarios with strict latency requirements in mobile edge networks.Based on the characteristics of partially shared and synthesized nearby user views,this thesis proposes a collaborative transmission and resource allocation algorithm based on hierarchical rendering and view synthesis technology.In addition,this thesis explores the allocation and optimization of computational and communication resources based on pre-rendering strategies when users move randomly.Specifically,the main contributions and innovative achievements of this paper can be divided into the following two parts.This thesis investigates the computing offloading problem of user view data in VR-based multi-user mobile edge networks and proposes a collaborative transmission and resource allocation strategy based on view synthesis to reduce the total amount of data that the server needs to compute.Firstly,this thesis models the physical and view locations of users and clusters them according to their channel correlation to distinguish users who need cooperation.Considering the mismatch of users’positions between the virtual world and the real world,surrounding users’views available for synthesis in the virtual world may be difficult to obtain in the real world.This thesis formulates this math problem between collaborating and collaborated users as an integer programming problem that maximizes the user’s virtual viewpoint distance increment.After solving this problem and obtaining the optimal strategy,this thesis formulates the transmission process of background views as a multiobjective non-convex optimization problem that maximizes the minimum data rate.This problem contains two stages.In the first stage,the base station transmits the shared background views to collaborating users.In the second stage,synthesized views are transmitted through the collaborative transmission channel assigned by the base station.To solve this problem,this thesis proposes an efficient iterative algorithm based on block coordinate descent and semi-definite relaxation.Simulation result shows that the proposed algorithm can effectively cluster and match users and improve the minimum transmission data rate for users.To address the issue of service interruption caused by user mobility in the multi-base and multi-user VR environment,this thesis proposes a prerendering-based cooperative transmission strategy for mobile VR users.This thesis models users by assuming that they may enter or leave the base station service range with a certain probability.For those who may enter the base station service range and those who always stay within the cell,the base station will pre-render the required view data.During the actual transmission stage,the base station can determine the actual users who enter or leave the base station service range based on the received signal power.Considering that pre-rendering view data for users who have not actually entered the base station service range will consume additional computing resources,this thesis derives a loss factor of prediction(LFP)and formulate the pre-rendering process as a problem of minimizing the LFP to optimize the allocation of computing resources.For users who have not actually left the base station service range,the surrounding users share the low-resolution transcoded data through cooperative transmission.The entire pre-rendering-based cooperative transmission process is modeled as a multi-variable problem to minimize the average delay,in which the base station supports both downlink visual data transmission and device-todevice transmission with frequency division multiplexing.To solve this problem,an iterative algorithm based on block coordinate descent and sequential convex optimization is proposed.Simulation result shows that the proposed scheme can reduce the average delay and improve the utilization efficiency of computing resources under different probabilities of users entering/leaving the base station service range.