基于粒子滤波步行长度预测的移动ad hoc网络路由算法  

A Mobile ad hoc Network Routing Algorithm Based on Walking Length Prediction after Particle Filtering

作　　者：张玲[1] 聂少华[2] 

机构地区：[1]北京信息职业技术学院电子工程系,北京100015 [2]临沂大学初等教育学院,山东临沂276000

出　　处：《电讯技术》2016年第3期331-336,共6页Telecommunication Engineering

摘　　要：针对移动ad hoc网络拓扑结构变化大、路由复杂度高、数据传输性能低等问题,提出了一种新的移动通信系统自适应路由算法。为了使得网络拓扑结构更接近移动网络间歇性连接的特点,该算法在网络结构上采用了一种改进的Levy Walk移动模型。采用一种粒子滤波步行长度预测的方法,通过蒙特卡罗抽样得到递归贝叶斯滤波器,并在粒子滤波后进行步行长度预测,确定消息的副本数量,从而减少由于节点转发过多消息副本带来的能量消耗量,提高消息的传递效率。实验仿真结果表明:与基于改进蚁群优化和利润优化模型的路由算法相比,该算法的消息传递成功率分别提高了0.08和0.04,节点平均能量效率提高了17.9%和13.4%,在提升数据传输成功率和节能上具有较好效果。For the problems of mobile ad hoc network( MANET) such as big change of topology structure, high routing complexity,and low data transmission performance,this paper proposes a new mobile commu-nication system using an adaptive routing algorithm. In order to make the network topology closer to the characteristics of the mobile network intermittent connection,an improved Levy Walk Mobility Model is a-dopted in the network structure. And then,a particle filtering walking length method prediction is used to get a recursive Bayesian filter throush Monte Carlo sampling,and the walking length is predicted after parti-cle filtering to determine the number of copies of message, thereby reducing the energy consumption for node forwarding a copy to bring too many messages and improve the message delivery efficiency. Simulation results show that compared with two routing algorithms of ant colony-based optimization and profit optimi-zation-based model,the proposed algorithm improves the message passing success rate of 0. 08 and 0. 04, node average energy efficiency of 17. 9% and 13. 4%,respectively. So it achieves better results in impro-ving the success rate of data transfer and saving energy.

关 键 词：移动ADHOC网络 粒子滤波 步行长度预测 LevyWalk移动模型 自适应路由算法 

分 类 号：TN915[电子电信—通信与信息系统]