The Research On Trajectory-Prediction-based Charging In Wireless Rechargeable Sensor Networks

With the development and popularization of Internet of Things technology,wireless sensor networks are widely used.In Wireless Sensor Networks(WSN),the energy problem plays the critical role in network performance and lifetime,because of the limited battery capacities of sensor nodes.Recently,Wireless Power Transmission(WPT)technologies provide a promising approach to address the energy problem in WSN.Researchers construct Wireless Rechargeable Sensor Networks(WRSN),which introduce mobile chargers(MC)with high capacity batteries for charging sensor nodes to prolong the lifetime of WSN.Recently,most studies in WRSN have paid attention to charging static nodes or mobile nodes with deterministic trajectories,whose real-time positions can be obtained directly.In these works,they can schedule the mobile chargers to charge sensor nodes at the positions on their travel paths.However,in the scenarios like wildlife tracking,where the sensor nodes are attached to the monitoring objects with non-deterministic mobility,even the mobility model itself is the target of the monitoring network,the charging position cannot be obtained directly.So the existed work could not work well in these scenarios.This thesis explores how to charge nodes with non-deterministic mobility.A novel approach,named Predicting-Scheduling-Tracking(PST),is proposed to perform charging tasks in these scenarios.Different from the existing work,the proposed scheme instructs the mobile charger to chase the sensor node and recharge it.Specifically,this thesis accomplishes the following contributions:1)The research background of the node energy problem in wireless sensor networks and the existing wireless energy transmission technologies are introduced,and the main structure and research status of wireless rechargeable sensor networks are analyzed and summarized.2)For the first time,an active tracking approach is proposed for charging sensor nodes with non-deterministic mobility,and the three main problems that need to be solved by using the active tracking approach are summarized and analyzed.3)Furthermore,the thesis proposes a Trajectory-Prediction-based charging algorithm named PST(Predicting-Scheduling-Tracking-based charging).In the PST algorithm,the base station periodically runs an improved LSTM to achieve the prediction of the next positions of the nodes.Based on prediction results and the current energy state of the nodes,the mobile charger runs the charging scheduling algorithm to select the most suitable sensor node in the network as the charging target and tries to meet the node to complete charging.In the wireless charging process,the mobile charger tracks the movement of the target node based on the node tracking algorithm to maintain the relative distance between the two sides of the wireless charging to meet the wireless charging requirements.4)A series of simulation experiments are conducted to verify the performance of the PST charging scheduling algorithm.The simulation results show that the PST algorithm can effectively guide the mobile charger to provide charging service for nodes with non-deterministic mobility and extend the lifetime of the WRSN.The comparison of charging efficiency metrics such as the number of depleted sensor nodes in the network,the average remaining energy of all sensor nodes,and the distance traveled by the mobile charger at the end of the simulation illustrates that the PST algorithm can maintain the wireless rechargeable sensor network in a better working condition compared to existing charging scheduling algorithms RLC(Reinforcement Learning-based charging).