Research On Indoor Moving Target Localization And Tracking Method

Dynamic location is one of the most important technologies in wireless sensor networks(WSN),which is widely used in many fields,such as digital agriculture,inspection of warehouse objects,underground exploration and so on.However,due to the complexity and variability of the experimental environment and the moving state of the unknown node,the success rate and accuracy of the node localization and tracking are seriously affected,especially the localization and tracking of the unknown node moving target in the indoor environment.Therefore,the research on indoor localization and tracking is of great significance.At present,most of the indoor localization and tracking algorithms still have shortcomings in the research of indoor environment and mobile target localization,which are mainly in the following aspects.The first is the impact of different obstacles on signal propagation in indoor environments.Secondly,most of the objects located and tracked are stationary nodes or moving nodes according to path planning.Thirdly,the accuracy of indoor localization and tracking needs to be improved.In view of the previous problems,this paper mainly improves from three aspects.The first is to establish a new signal loss model which is more suitable for the experimental environment.The second is to improve the localization algorithm to increase positioning accuracy and positioning success rate.The third is to improve the tracking and filtering algorithm to make the tracking path of the unknown node closer to the real path and achieve path optimization.Model building stage,this paper proposes multi-structure signal loss model through a large number of experiments in the indoor environment.According to the measured data,data curve fitting was carried out on the corresponding software,and finally the parameters in the signal loss model were obtained.There are two main factors affecting the signal strength.They are the distance between the two nodes and the obstacles in the process of signal propagation.Field measurement is also divided into two categories.The first is the relationship between different distances and signal strength under the same condition of obstacles that affect signal propagation.The second is the relationship between different obstacles and signal strength under the condition of the same distance.The improvement stage of localization algorithm is based on the model proposed in the model establishment stage.This paper proposes the Indoor moving target localization algorithm based on WSN region division(RDLA).According to the serial numbers and coordinates of the reference nodes,the experimental areas are divided into three categories.They are room,corridor and stair space.Thus,a unique coordinate range can be determined at each registration point.At each registration point,the unknown node is positioned with its coordinate range as the limiting condition.The unknown node was not successfully located or the location coordinates were not in the range of coordinates at some registration points.In order to solve this problem,the center of mass of intersection space is used as the coordinate of the locating point.In the tracking algorithm improvement stage based on the coordinates located in the positioning algorithm improvement stage,a markov-kalman filter tracking algorithm for indoor moving targets(MKF)is proposed.The algorithm firstly forms the sampling points around the positioning coordinates by using the nonlinear formula according to the markov properties and calculates the weights of each sampling point.The sampling points are added as filter objects to kalman filter,and their respective weights are added.Finally,the moving path of the unknown node is obtained.The simulation results show that compared with the trackless kalman tracking algorithm and the adaptive Markov matrix IMM tracking algorithm,the average tracking error of MKF algorithm decreased by 1.216% and 0.71%,respectively.