一种基于地理位置预测的高性能路由算法  

High Performance Routing Algorithm Based on Geographic Location Prediction

作　　者：沙毅[1] 郭自强[1] 朱丽春[2] 张志伟[2] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819 [2]中国科学院国家天文台,北京100012

出　　处：《东北大学学报（自然科学版）》2015年第9期1242-1245,共4页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(11273001)

摘　　要：提出了基于 ARIMA 预测模型的高效路由算法。该算法中节点通过前向与反向成功转发率、数据传输速率等计算链路的丢包率和期望传输次数来获取干扰感知期望传输时间（iETT），代替 DSR 路由算法中的最短跳数判据。并引入 ARIMA 模型来预测节点下一时刻的运动位置，防止链路频繁断裂造成的网络丢包，并在链路失效之前预先选择最稳定的路径进行数据传输。仿真结果表明，所提路由算法相比 DSR 判据吞吐量提高6%~9%，平均端到端时延降低2%~6%，提高了网络整体性能。Based on the ARIMA prediction model,an efficient routing algorithm was put forward.In this algorithm,a node could obtain the interference-aware expected transmission time （iETT）by measuring the packet loss rate and expected transmission number with the successful forward and reverse forwarding rate as well as the speed of data transmission.Thus,the L-iETT could replace the shortest hop criterion of the DSR routing algorithm.The ARIMA model was introduced to predict the movement location of the next node so as to prevent the network packet loss caused by frequent link fractures and select in advance the most stable path to data transmission before link failure.The simulation results showed that the routing algorithm criterion increases by 6% ~9% compared with the DSR throughput and the average end-to-end delay reduces by 2% ~6% so that the network＆#39;s overall performance can be improved.

关 键 词：地理位置预测 iETT 路由判据 多速率 AD HOC 网络 

分 类 号：TP393.17[自动化与计算机技术—计算机应用技术]