The support of coexisting ultra-reliable and low-latency (URLL) and enhanced Mobile BroadBand (eMBB) services is a key challenge for the current and future wireless communication networks. Those two types of services introduce strict, and in some time conflicting, resource allocation requirements that may result in a power-struggle between reliability, latency, and resource utilization in wireless networks. The difficulty in addressing that challenge could be traced back to the predominant reactive approach in allocating the wireless resources. This allocation operation is carried out based on received service requests and global network statistics, which may not incorporate a sense of \textit{proaction}. Therefore, this paper proposes a novel framework termed \textit{service identification} to develop novel proactive resource allocation algorithms. The developed framework is based on visual data (captured for example by RGB cameras) and deep learning (e.g., deep neural networks). The ultimate objective of this framework is to equip future wireless networks with the ability to analyze user behavior, anticipate incoming services, and perform proactive resource allocation. To demonstrate the potential of the proposed framework, a wireless network scenario with two coexisting URLL and eMBB services is considered, and two deep learning algorithms are designed to utilize RGB video frames and predict incoming service type and its request time. An evaluation dataset based on the considered scenario is developed and used to evaluate the performance of the two algorithms. The results confirm the anticipated value of proaction to wireless networks; the proposed models enable efficient network performance ensuring more than $85\%$ utilization of the network resources at $\sim 98\%$ reliability. This highlights a promising direction for the future vision-aided wireless communication networks.
翻译:支持超可靠和低延迟共存(URLL)和增强的移动宽带(EMBB)服务是当前和未来的无线通信网络面临的一个关键挑战。这两种服务类型引入了严格的资源分配要求,有时会相互冲突,可能导致无线网络的可靠性、延时率和资源利用之间形成电源结构。应对这一挑战的困难可追溯到分配无线资源的主要被动反应方法。这一分配业务是根据收到的服务请求和全球网络统计数据进行的,这可能不会包含某种 Textit{proaction}感。因此,本文件提出了一个名为\ textit{sidation}的服务识别的新框架,以开发新的积极主动的资源分配算法。 开发的框架基于视觉数据(例如RGB相机的描述)和深度学习(例如深神经网络 ) 。 这个框架的最终目标是使未来的无线网络具备分析用户行为模式、预测即将到的服务以及积极主动的资源分配能力。 展示拟议框架的潜力,一个名为\ textrealrea Netwo 的网络运行方向,一个在使用基于预想到的 RGB URL 和 REMB 的预估值的预估测的预估测的 RURL 数据框架上,一个基于 URL URL 和 的预测的预测的预算的预测算的预算的网络的预估测算的预算 和预算,是利用基于的 RURL 和预估测算的 RURL 的预算的预算的 RURMB 的网络的预算的 RURL 的 RURL 的 RL 和深的 RL 的预算的预算的 RURL 和深的预算的 RURB 的逻辑的预算的 RL