Research On Indoor Pedestrian Localization Enhancement Technology Based On Inertial Navigation

With the development of The Times,precise navigation and positioning has become more and more urgent in People’s Daily life,production services,emergency rescue and other fields.Currently,the Global Position System(GPS)and Beidou navigation satellite System,which are widely used,are mainly aimed at outdoor positioning.Because they are sheltered by buildings indoors,their signals decay rapidly and cannot meet the needs of indoor positioning.Therefore,the Inertial Measurement Unit(IMU)as the core of the inertial navigation system,because of the characteristics of free from infrastructure limitations,small size,strong portability and so on,has become the research hotspot of indoor autonomous positioning.However,due to the limited accuracy of the inertial measurement unit and the long-term work,the noise of the inertial sensor itself accumulates for many times,which leads to the rapid divergence of the pedestrian track and affects the positioning accuracy.Doing research on the inertial navigation technology of indoor pedestrian independent positioning enhancement technology,this thesis proposes a new optimal threshold estimation algorithm to reduce the integral cumulative error,and on this basis,reduces the pedestrian heading drift through the map matching based on the improved particle filter type,realizes the improvement of the indoor pedestrian positioning accuracy.(1)Based on the positioning technology of installing IMU only at the foot of pedestrians and the Simultaneous Location And Mapping(SLAM)method,the Foot SLAM algorithm is used to achieve the optimal estimation of pedestrian trajectories in an unknown environment.In the zero-speed update stage,a generalized likelihood test is used to detect the zero-speed update points and Kalman smoothing algorithm is used to filter the predicted value of the system state,which makes the noise suppression effect better and the pedestrian position more accurate.(2)A new method of optimal threshold estimation using adaptive threshold and Foot SLAM matching is proposed,and the pedestrian location error is compared in the fixed threshold method alone or the adaptive threshold method based on speed classification.The experiment verifies the effectiveness of the new method of indoor pedestrian location enhancement.Compared with the fixed threshold method and the adaptive threshold method based on speed classification,the average horizontal position error of the new method is reduced by more than 80%.Moreover,when all the data are connected into a single trace,compared with the adaptive threshold method based on speed classification,the adaptive threshold method proposed in this thesis can reduce the norm of horizontal position error to10.1%.Compared with the optimal fixed threshold method,the adaptive threshold method proposed in this thesis can reduce the horizontal position error norm to 25.6%.(3)On the basis of the new method of threshold optimal estimation,the pedestrian location is realized by improved map matching based on particle filter.There are two specific improvements: first,for the problem of particle passing through the wall,the inaccessible or not directly accessible areas are considered into the constraint particle in the particle weight update;The second is the correction of particles in special geographical locations.Through the experimental test,the effectiveness of the proposed indoor pedestrian positioning enhancement technology is verified,which can effectively improve the accuracy of indoor pedestrian positioning.The pedestrian location error is controlled within 4.3% by the optimal threshold estimation and the improved map matching method based on particle filter.