Research On Mobile Augmented Reality Based On MEC

Augmented Reality (AR) refers to the computer-generated virtual information superimposed on the real scene, to achieve the effect of information enhancement. It can enhance the user's perception of the real world and greatly improve the user's true sense of experience. As smartphone performance improves, developing augmented reality applications on the mobile device becomes possible. Mobile AR applications have huge application potential in commercial, entertainment,engineering, education and other fields. So mobile AR has aroused extensive attention and research of scholars. Recognition of real-world scenarios requires the use of computer vision algorithms for complex operations, as well as search queries in large sample libraries.Overlapping virtual-enhanced content for a specific scenario,involves real-time push and content updates. Because the computing and storage resources of mobile phones are limited, these features require mobile devices and servers to collaborate,so the availability,performance and popularity of mobile augmented reality applications depend largely on the overall design of the system and the network environment . From the network perspective, Mobile Edge Computing (MEC) is an effective way to support real-time, high-bandwidth, near-local and localized emerging services like augmented reality.In view of this, to achieve a better performance of MobAR applications, this paper proposes a MobAR innovative application system solutions based on MEC. The MobAR application is ported to the MEC platform, part of the calculation and storage functions is transferred to the MEC server. MEC traffic offload, low latency, high bandwidth and location awareness enhance the performance of MobAR applications.This paper realizes the design scheme, and presents the augmented reality effect on campus. And the performance of the system was evaluated by several experiments. Specifically include the following:First of all, this paper designs and implements the MEC-based MobAR application system composed of MobAR client, MobAR server and user information acquisition module. MobAR client provides real-time video digital acquisition, feature detection and description,image rendering, human-computer interaction and other functions. The MobAR server is deployed on the MEC server to provide storage and retrieval of AR target samples, and storage and delivery of AR content and so on. The user information acquisition module provides wireless network access and knows the user location.Secondly, the functional verification and performance evaluation of the application system are carried out. The MobAR system achieves a augmented reality result of superimposed text, pictures, videos, charts other in Beijing University of Posts and Telecommunications campus sculpture, buildings, signs and other scenes.In aspect of performance evaluation, the paper compares and analyzes the application results in MEC and mobile cloud computing network environment. After testing the performance of the public cloud server network, the MCC network environment was set up by implementing the network simulator. The different time delay, packet loss and other time parameters used during the application were evaluated. Finally, the MEC-based MobAR application was proved to be superior sexual and has practical value.