无线网络分层QoS垂直映射模型及跨层优化方法  

Layered QoS Vertical Mapping Model and Cross Layer Optimization for Wireless Networks

作　　者：房曙光[1,2] 唐日照[2] 董育宁[1] 

机构地区：[1]南京邮电大学通信与信息工程学院,江苏南京210003 [2]无锡商业职业技术学院物联网技术学院,江苏无锡214153

出　　处：《南京邮电大学学报（自然科学版）》2014年第6期34-40,共7页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家自然科学基金(60972038;61271233;61101105;61001077);教育部博士点基金(20103223110001);江苏省2011年度普通高校研究生科研创新计划(CXZZ11_0392);江苏省高等教育自然基金(10KJB510013);南京邮电大学科研基金(NY208052)资助项目

摘　　要：现代无线通信网络结构复杂,通信环境时变,业务流和无线信道不可预知,服务质量保障极具挑战性。基于无线网络分层QoS体系结构,文中提出了一种统一的基于协议层间虚拟缓存接口QoS垂直映射模型,以虚拟队列技术实现层间QoS指标映射。同时,基于该协议层间QoS垂直映射模型提出一种跨层QoS优化架构,并针对业务流和无线信道的不可预知性,给出该架构在线学习优化求解方法。最后,通过与现有文献技术对比一致性方法验证所提出层间QoS垂直映射模型正确性,并且在物理层包差错率小于30%条件下,分析Q学习法优化数据链路层时延收敛性,从而验证所提出跨层优化模型和在线学习优化方法的有效性。The complex structure and time varying environment in modern wireless communications make traffic flow and wireless channel be unpredictable, which arises more challenges for QoS guaranteeing. Based on the layered QoS system architecture, this paper proposes a unified layered QoS vertical mapping model depending on the virtual buffer interface between protocol layers. In addition, an online learning optimization algorithm is presented for the model due to the unpredictable traffic flow and wireless channel status. The efficiency is evaluated by comparing with the previous work. Furthermore, under the scenarios of packet error rate less than 30% for physical layer, the convergence of delay optimizing is analyzed in layer,thus verifying the proposed cross-layer QoS optimization framework and Q-learning optimizing method.

关 键 词：无线网络 垂直QoS映射 跨层优化 在线学习 

分 类 号：TN915.0[电子电信—通信与信息系统]