| In recent years,Wireless Sensor Networks(WSN)are widely used in military,intelligent transportation,human health monitoring and other fields.These application scenarios require WSN to work continuously for a long time period.However,due to the small size of sensor nodes and the characteristics of battery power supply,the energy carried by sensor nodes is limited,so it is necessary to supplement energy for sensor nodes through charging equipment to extend the network life.Compared with fixed charging device,mobile charging device(Wireless Charging Equipment,WCE)has higher flexibility and controllability;compared with full coverage mobile charging strategy,ondemand charging strategy based on sensor node charging request is more practical.Therefore,the on-demand charging strategy of WSN has become a hot topic for scholars.However,most of the existing charging strategies use greedy algorithms to determine the charging path for WCE,which limits the performance of charging strategies.Moreover,the existing strategies ignore the wireless charging characteristic of WCE,which results in the low charging efficiency of WCE.To solve the above problems and improve the performance of WSN and WCE at the same time,this thesis studies the mobile charging strategy of WSN when the velocity of WCE is variable,mainly including the following two aspects:(1)An on-demand charging planning algorithm based on swarm reinforcement learning is designed,and the WCE can independently plan its charging path.The ondemand charging problem is mapped to reinforcement learning.Through the discretization of state space and behavior space of WCE,the on-demand charging planning problem is modeled as reinforcement learning problem,and the reward model is designed based on the energy of sensor network and the energy utilization of WCE under the limited energy.Then simulation results show that the proposed algorithm can effectively extend the network life and improve the energy utilization of WCE.(2)The velocity control algorithm of WCE in wireless sensor networks is proposed.Firstly,considering the relationship between charging power and distance,different charging areas are divided from the view of sensor nodes.Then the charging path is divided into several segments according to the charging area.Next,by controlling the velocity of WCE on the segment,under the condition that WCE is fully charged for sensor nodes,the charging completion time of WCE is reduced.And the station time is increased,so as to improve the charging efficiency of WCE.Finally,the effectiveness of the proposed algorithm is verified by simulation experiments. |