基于MDP的分簇水声网络TDMA协议

A TDMA Protocol for Clustered Underwater Acoustic Networks Based on Markov Decision Process

TDMA协议是固定分配时隙的MAC协议,由于水声信道具有传播延迟高、带宽窄等特点,水声网络协议的效率较低。本文针对分簇网络结构提出一种基于马尔可夫决策过程(MDP)的动态分配时隙的MDP-TDMA协议,该协议根据簇成员节点的本地信息来计算下一轮所需要的时隙数,减少了成员节点与簇头节点的信息交换,节省了信道资源的同时减轻了簇头节点的负担。仿真实验结果表明,与传统TDMA协议相比,MDP-TDMA协议在网络生存周期、网络吞吐量及丢包率等性能上得到了提升。

TDMA is the MAC protocol based on a fixed allocation. Due to the high propagation delay and narrow bandwidth of the underwater acoustic channel, the efficiency of the underwater acoustic network protocol is low. In this paper, a MDP-TDMA protocol based on Markov Decision Process(MDP) is proposed for the clustered network structure. The protocol calculates the number of time slots required for the next round based on the local information of the cluster members, which decreases the information exchange between the member nodes and the cluster head node, saves the channel resources and reduces the burden of the cluster head node. The simulation results show that compared with the traditional TDMA protocol, the performance of MDP-TDMA protocol has been improved in terms of network lifetime, network throughput and packet loss rate.