As the essential technical support for Metaverse, Mobile Augmented Reality (MAR) has attracted the attention of many researchers. MAR applications rely on real-time processing of visual and audio data, and thus those heavy workloads can quickly drain the battery of a mobile device. To address such problem, edge-based solutions have appeared for handling some tasks that require more computing power. However, such strategies introduce a new trade-off: reducing the network latency and overall energy consumption requires limiting the size of the data sent to the edge server, which, in turn, results in lower accuracy. In this paper, we design an edge-based MAR system and propose a mathematical model to describe it and analyze the trade-off between latency, accuracy, server resources allocation and energy consumption. Furthermore, an algorithm named LEAO is proposed to solve this problem. We evaluate the performance of the LEAO and other related algorithms across various simulation scenarios. The results demonstrate the superiority of the LEAO algorithm. Finally, our work provides insight into optimization problem in edge-based MAR system for Metaverse.

翻译：作为元宇宙的关键技术支撑，移动增强现实（MAR）已引起众多研究者的关注。MAR应用依赖于视觉和音频数据的实时处理，这些繁重的计算任务会迅速耗尽移动设备的电池电量。为解决该问题，基于边缘的解决方案应运而生，用于处理需要更高计算能力的任务。然而，此类策略引入了新的权衡：降低网络延迟和整体能耗需要限制发送至边缘服务器的数据量，而这将导致精度下降。本文设计了一种基于边缘的MAR系统，并提出数学模型对其进行描述与分析，揭示延迟、精度、服务器资源分配与能耗之间的权衡关系。此外，提出了一种名为LEAO的算法用于解决该优化问题。我们通过多种仿真场景对LEAO及其他相关算法的性能进行评估，结果表明LEAO算法具有优越性。最后，本研究为元宇宙中基于边缘的MAR系统的优化问题提供了深入见解。