| With the development of science and technology,most of the people’s social production and living activities are carried out indoors,and the demand for indoor location services is increasingly urgent.Due to signal masking or attenuation,the existing satellite navigation and positioning systems are difficult to provide reliable positioning information in an indoor environment.Therefore,high-precision indoor navigation and positioning technology has attracted much attention in recent years and has become the key to the application of various indoor location services.Inertial measurement is not affected by time,place,or environment,and can realize positioning and navigation in the indoor environment,which has become one of the research hotspots of indoor positioning.However,inertial measurement has problems such as integral accumulation error and course drift error,which greatly affect the improvement of indoor positioning accuracy.This thesis explores the inherent error reduction method in inertial measurement and provides technical support for achieving high precision indoor positioning based on the inertial measurement unit.The main work of this thesis is as follows:(1)This thesis deeply studies the solution of inertial navigation integral cumulative error and analyzes the problems existing in the zero velocity update algorithm.Most of the traditional zero-speed detection algorithms are based on a fixed threshold or an extended one,which has poor robustness.To improve the accuracy of the indoor pedestrian inertial navigation system,this thesis proposes a zero-speed detection algorithm based on AttentionLSTM.It does not need to change the threshold,and can accurately detect zero-speed points for movements at different speeds,which improves the robustness and accuracy of zerospeed detection.The experimental results show that under different speeds of 1.1m/s,2.2m/s,and 3.3m/s,the inertial navigation distance error of the zero-speed detection algorithm based on attention-LSTM is 4.75 m on average,and its performance is far better than other traditional algorithms.(2)Although the zero velocity update algorithm can suppress the accumulation error growth,it cannot effectively correct the course.In this thesis,map matching is used to correct the trajectory of pedestrian drift.Aiming at the problem that the traditional map matching algorithm is suitable for a single scene,this thesis proposes a map matching method based on Wi Fi assistance taking advantage of the different physical conditions of indoor corridors and rooms,which enables map matching by judging whether the pedestrian is in the corridor or the room.Experimental results show that the average starting point error of the map matching solution based on Wi Fi is 0.894 m.The proposed algorithm can better correct the drifting trajectory,and the positioning effect is excellent.(3)Aiming at the trajectory drift of pedestrians in the room,this thesis introduces a Wi Fi fingerprint positioning method based on the Wi Fi-assisted map matching,which uses an unscented Kalman filter to fuse PDR and Wi Fi positioning information based on Ada Boost.The accuracy of pedestrian positioning in the room got improved,as well as the track through the wall was corrected after the complementary advantages of the two technologies.The experimental results show that the average distance error of the proposed navigation and positioning system is 1.489 m,and the indoor pedestrian autonomous navigation and positioning system has a good effect. |
PDF Full Text Request