Research On Node Localization And Target Tracking In Large-scale Wireless Sensor Network

Large-scale wireless sensor networks(WSN)often consist of hundreds of sensor nodes,which are widely used in scenarios such as wildlife monitoring,intelligent traffic control,and intrusion target detection in key areas,where target localization and tracking are key issues that need to be studied.Nodes in large-scale WSNs are often deployed in a randomly thrown manner,and carrying positioning devices on all nodes is costly and energy-intensive,while the location information of nodes is the basis for achieving target localization and tracking;after the target enters the monitoring area,due to the dense deployment of sensor nodes and overlapping coverage,waking up too many nodes to perform tracking tasks will lead to problems such as redundancy of sensory information and increased energy consumption.The balance of tracking accuracy and node energy consumption is affected by waking up a small number of nodes or fixed nodes to perform tracking tasks.Therefore,in thesis,the following research is conducted to address the problems of autonomous positioning of sensor nodes,target localization and cooperative tracking of nodes involved in target localization and tracking in large-scale WSN(1)For the scenario where only a small number of anchor nodes with localization capability exist in a large-scale WSN,a comprehensive and improved distributed collaborative node localization algorithm is proposed in thesis.By analyzing the sources of localization error in each stage of the traditional DV-Hop(Distance Vector Hop)localization algorithm,mechanisms such as RSSI hop quantization,anchor node screening and collaboration,and differential evolution algorithm are introduced to optimize the calculation of its different stages.Simulations show that the localization accuracy of this algorithm is improved by 57%and 38%over the original algorithm and the improved algorithm in literature^[44]under the condition that the percentage of anchor nodes is only 5%,and the advantage is more obvious in large-scale WSNs.(2)For large-scale homogeneous WSNs,a dynamic cluster-based target localization and distributed node cooperative tracking algorithm is proposed in thesis.The target is tracked by local collaboration among neighboring nodes to form a tracking cluster,and the monitoring range,number,and sensing period of wakeup nodes within the cluster can be adaptively adjusted according to the target state to balance the tracking accuracy and energy consumption.In addition,a two-stage target state estimation algorithm based on the unscented kalman filter is designed to improve the accuracy of the target localization and nonlinear tracking system.The simulation shows that under the same scenario,the tracking error of this algorithm is reduced by 59%compared with the original dynamic cluster tracking algorithm,and the energy consumption is only increased by 7%.The algorithm can effectively reduce the number of target tracking loss,and can recover tracking with lower system consumption after the target is lost.(3)For large-scale heterogeneous WSNs,a centralized target tracking and node scheduling algorithm based on multi-agent collaboration is proposed in thesis,taking into account the inability of direct communication between heterogeneous nodes and the differences in sensing capabilities.The agent acted by the aggregation nodes implement collaborative scheduling by sharing value functions among them and combining the neighbor information,thus making the tracking and scheduling decisions globally optimal.The simulation shows that under the same scenario,compared with the traditional greedy algorithm and the single-agent algorithm,the tracking accuracy of the algorithm is increased by 68%and 36%,and the system energy consumption is reduced by 13%and15%,respectively,and the tracking stability of the algorithm is better.