Research On Joint Positioning Algorithm Of Personnel Target Based On Multi-sensor Information Fusion

In an informationized battlefield environment,timely and accurate acquisition of the location information of enemy personnel targets is of great significance to tactical design and strategic decision-making.Passive positioning technology achieves positioning by analyzing and processing signals emitted by enemy targets.It has good concealment and a wide positioning range.Therefore,it is highly concerned by countries around the world and widely used in modern electronic warfare.This paper takes the land battlefield as the background,uses vibration sensors and vision sensors as detection equipment,and uses TDOA positioning algorithm,AOA positioning algorithm,multi-target data association algorithm and fusion positioning algorithm as the main processing algorithms to study the joint positioning algorithm of personnel targets based on multi-sensor information fusion,and build a complete usable software and hardware system.The specific research work is as follows:(1)Research vibration sensor positioning algorithm based on TDOA and vision sensor positioning algorithm based on AOA.For vibration positioning,the time delay estimation algorithm is used to calculate the time delay value of the vibration signal to each sensor,and the seismic wave propagation speed is used to construct the TDOA equation system.The Chan algorithm,Taylor algorithm,or AML algorithm is used to obtain the position estimate of the target.For vision positioning,the YOLOv3 target detection algorithm is combined with the similar triangle principle of the small hole imaging model to extract the azimuth of the target to construct the AOA equation system,and the target position estimate is obtained through the TLS algorithm or the TLS-KF algorithm.Finally,using the root-mean-square error of the positioning result as the evaluation criterion,the time delay estimation algorithm and positioning algorithm used in the joint positioning system are determined through simulation experiments.(2)Aiming at the application scenarios of multi-sensor and multi-target,study the measurement-target association of vibration positioning,the target-target association of vision positioning,and the association of the positioning results of the two.The measurement-target association problem of vibration location can be transformed into a multi-dimensional assignment problem to be solved.The Lagrange relaxation algorithm is used to reduce the dimension to a series of two-dimensional assignment problems that can be solved within a certain time complexity,and the Hungarian algorithm is used to solve it.For vision positioning,since the correspondence between measurement and target under the same sensor is known,the association problem is to find the matching of each target under different sensors,which can be converted into a pedestrian re-identification problem for processing.Through feature extraction,distance measurement and reordering complete the target matching.In order to improve the accuracy of association,considering the angle measurement and appearance characteristics of the target,this paper proposes a DSm T-based multi-distance fusion association algorithm.Experiments prove that it has a higher association accuracy than the pedestrian re-identification technology alone.Finally,the nearest neighbor algorithm is used to quickly realize the association of the two positioning results.(3)In order to further improve the positioning accuracy,the decision-level fusion positioning algorithm for heterogeneous sensors is studied.In decision-level fusion positioning,the vibration sensor and the vision sensor respectively use the TDOA positioning algorithm and the AOA positioning algorithm for position estimation,and then fuse the estimation results according to the principle of minimum mean square error or best linear unbiased principle.Finally,the root-mean-square error of the positioning results is also used as the evaluation standard,and the fusion positioning algorithm used in the joint positioning system is determined through simulation experiments.Finally,according to the overall positioning process,build a hardware and software system architecture,and design the corresponding GUI program to form a complete and usable joint positioning system,and conduct field testing and experimental analysis of each link.