基于链路质量与节点负载估计的Q学习UANET路由协议

Q-Learning UANET Routing Protocol Based on Link Quality and Node Load Estimation

针对无人机自组网(UANET)网络拓扑变化频繁,传统路由协议建立链路的稳定性较差,而导致的链路断裂、高负载情况下的业务丢失等问题,提出了基于链路质量与节点负载估计的Q学习UANET路由协议,该协议在最优链路状态路由(OLSR)协议的基础上,使用Q-learning算法,将跳数、链路质量和节点负载作为路由决策的奖励函数,以加权的综合评价指标作为路由的决策准则。实验结果表明,改进后的协议更加适用于无人机自组网中高动态、高负载的网络环境。

Due to the frequent changes in the network topology of unmanned aerial vehicles Ad hoc networks(UANET),the poor stability of the links established by traditional routing protocols leads to the link breakage,service loss under high-loading scenario,and other problems.Therefore,a Q-learning UANET routing protocol is proposed based on the link quality and node load estimation.On the basis of the optimized link state routing(OLSR)protocol,this protocol adopts the Q-learning algorithm to use the hop count,link quality,and node load as reward functions for routing decisions,and a weighted comprehensive evaluation indicator is adopted to be the decision-making criteria for routing.The experimental results show that the improved protocol is more suitable for the network environments with high dynamic and high load.