摘要
针对传统无线传感器网络节点能量供应有限和网络寿命短的瓶颈问题,依据无线能量传输技术领域的最新成果,提出了一种基于改进Q-Learning的无线可充电传感器网络的充电路径规划算法。基站根据网络内各节点能耗信息进行充电任务调度,之后对路径规划问题进行数学建模和目标约束条件设置,将移动充电车抽象为一个智能体(Agent),确定其状态集和动作集,合理改进ε-greedy策略进行动作选择,并选择相关性能参数设计奖赏函数,最后通过迭代学习不断探索状态空间环境,自适应得到最优充电路径。仿真结果证明:该充电路径规划算法能够快速收敛,且与同类型经典算法相比,改进的Q-Learning充电算法在网络寿命、节点平均充电次数和能量利用率等方面具有一定优势。
Aiming at the bottleneck problems of traditional Wireless Sensor Network(WSN) nodes like limited energy supply and short network life, based on the latest achievements in the field of wireless energy transmission technology, a charging path planning algorithm based on improved QLearning Wireless Rechargeable Sensor Network(WRSN) is proposed. Firstly, the base station performs charging task scheduling based on the energy consumption information of each node in the network;and then mathematical modeling and target constraint setting are performed on the path planning problem.The mobile charging vehicle is abstracted as an agent, and its state set and action set are determined.The ε-greedy strategy is reasonably improved for action selection, and the relevant performance parameters are selected to design the reward function. Finally, the state space environment is explored through iterative learning to adaptively obtain the optimal charging path. The simulation results prove that the charging path planning algorithm can quickly converge, and has certain advantages in terms of network life, average charging times of nodes and energy utilization compared with the classic algorithms of the same type.
作者
刘洋
王军
吴云鹏
LIU Yang;WANG Jun;WU Yunpeng(School of Electronics and Information Engineering,Suzhou University of Science and Technology,Suzhou Jiangsu 215009,China;Changchun Institute of Optics,Fine Mechanics and Physics,Chinese Academy of Science,Changchun Jilin 130033,China)
出处
《太赫兹科学与电子信息学报》
2022年第4期393-401,共9页
Journal of Terahertz Science and Electronic Information Technology
基金
江苏省研究生科研创新资助项目(KYCX17_2060)
近地面探测技术重点实验室资助项目(TCGZ2018A005)
