The traditional single-Sink sensor networks have some disadvantages, such as consuming energy of the nodes on the key path too quick,the singleness of routing algorithm,the invalidation of the Sink node,etc. While multi-Sink WSNs can solve these problems efficiently,protocols of multi-Sink WSNs are studied and a new routing mechanism based on Q-learning is presented. By using Q-learning, nodes take various kinds of conditions into account, study periodically and achieve their optimal routing. Finally the source node chooses the optimal path to send signals according to the Q value. Case study shows that this mechanism saves the node energy,prolongs the life of sensor networks.%单Sink节点无线传感网存在部分关键区域节点能量消耗过快、路由选择算法单一及Sink节点失效的问题.多Sink节点无线传感网能有效解决上述问题,研究多Sink节点传感网的路由协议及其存在的不足,提出了一种基于Q学习的路由选择机制.利用该学习方法节点综合考虑各种环境因素,进行周期性的学习训练,优化路径选择,最后按照计算得到的Q评估值选择最优的路径进行数据传送.实例分析表明,基于Q学习的路由机制能有效延长网络的使用寿命.
展开▼