| Individual navigation equipment plays an extremely important role in military operations,rescue and disaster relief,and special operations.Accurate navigation information can greatly assist individual soldiers in completing combat tasks and guarantee their lives.Individual soldier operation scenarios are complex and changeable,and satellite positioning alone cannot provide services.Therefore,inertial sensing and its combined system have become the most important technical implementation of individual soldier navigation.Due to the non-linear zero drift of the inertial device,the installation error and the inconsistency of the performance of the inertial unit,and the large differences in the behavior of individual soldiers,it is difficult to maintain the stability of individual navigation accuracy.Artificial intelligence is the use of computer to simulate and expand human intelligence.As a new technology,it can fit and compensate for the nonlinear data in the components,units and system integration of individual navigation equipment,and become an indoor and outdoor individual.The hot spots in the research field of navigation accuracy optimization have important significance and value.Due to the autonomy and continuity of the inertial positioning technology,it can meet the requirements of indoor and outdoor individual navigation and positioning,but its positioning accuracy decreases with the accumulation of time.Therefore,this article is based on the optimization of the Pedestrian Dead Reckoning(PDR)algorithm,combined with GPS positioning and Wi-Fi positioning,and uses artificial intelligence technology to optimize the accuracy of indoor and outdoor individual navigation,and finally achieve improvement.The purpose of individual navigation system trajectory reduction and long-term stability.The research contents of this thesis mainly include:1.Research and improve the PDR positioning algorithm.Aiming at the problem of inaccurate step detection,a step detection method based on decision tree is proposed;for the problem of inaccurate step estimation,the four existing step estimation models are analyzed and compared,and artificial intelligence models are used.Step length estimation model of RBF;In view of the error caused by pedestrian body jitter in the course of heading estimation,the wavelet analysis algorithm is used to eliminate the error caused by pedestrian body jitter,and then combined with different walking modes,different heading calculation methods are adopted.2.Researched Wi-Fi fingerprint recognition and positioning.This technology includes two parts: offline database construction and online positioning.First,it analyzes the Wi-Fi signal distribution,then selects Gaussian filtering and mean filtering to filter it,then combines geographic coordinates to build a fingerprint database,and finally uses the traditional weighted K-nearest neighbor algorithm to achieve matching and positioning of the fingerprint database.3.This thesis proposes a SA-BP-based indoor and outdoor individual navigation accuracy optimization algorithm.Through the research of BP neural network and simulated annealing algorithm,an algorithm model for the accuracy optimization of indoor and outdoor individual navigation based on SA-BP is constructed.Through experiments,it is concluded that the indoor and outdoor individual navigation accuracy optimization algorithm based on SA-BP can optimize the accuracy of indoor and outdoor individual navigation,and effectively improve the trajectory restoration and long-term stability. |
