基于流量预测的无线mesh网络负载均衡路由协议

提出了一种基于神经网络预测模型的无线mesh网络负载均衡协议NNP-L2MPM。协议根据网络中泛洪的HELLO包计算路径质量,从而选择出到达目的节点的最优下一跳,并以MAC层接口队列长度作为流量负载的衡量依据,然后利用RBF神经网络预测模型对mesh网路中的节点流量负载进行预测,根据预测的下一时刻的流量负载优化路径质量,提前实现路由更新,避免中间节点发生拥塞,进而提高网络性能。仿真结果表明:与原有路由协议相比,所提协议在数据包投递率上提高了约9%,平均端到端延时降低了约16%。

Load Balancing Routing Protocol Based on Traffic Prediction for Wireless Mesh Networks

We proposed a load balancing routing protocol in the wireless mesh network based on neural network prediction model,namely NNP-L2MPM.In this protocol,the optimal next hop to the destination is selected by using the path quality which is calculated by HELLO packets flooded on the network.The traffic load is measured by the length of the queue at the interface in MAC layer.Then the RBF neural network model is used to forecast the traffic load in the nodes of the mesh network.According to the forecasted traffic load in the next time,path quality is optimized,routing is updated previously,congestion in the intermediate nodes is avoided,and finally network performance is improved.Compared with the original routing protocol,the results of simulation show that the rate of the packet delivery can increase about 9%,and the average end-to-end delay can reduce about 16%.