Key Technology And Applications Of Indoor Position And Navigation Based On Multi-sensors Fusion

Inertial Measurement Units(IMUs)have the advantages of small size,high accuracy and low power consumption,and are widely used in indoor and outdoor navigation.However,the shortcomings of sensor bias and integration accumulation error inherent to inertial navigation systems have not been well resolved,limiting the use of inertial navigation systems,especially in indoor complex pedestrian navigation and positioning,eliminating the sensor bias and integration accumulation error inherent to inertial navigation systems has become a bottleneck,and the study of high-precision indoor positioning and navigation systems has become the frontier in this field.This thesis explores the multi-sensor fusion indoor positioning and navigation technology based on inertial navigation technology.(1)To address the problem of poor robustness of the fixed-threshold zero velocity detection algorithm,this thesis discusses in detail the adaptive threshold zero velocity detection algorithm based on Bayesian filtering;to address the problem that the traditional zero velocity correction algorithm cannot correct the navigation heading,this thesis integrates zero angular velocity update and heuristic heading correction into the zero velocity update algorithm based on extended Kalman filter,and constructs the zero velocity correction algorithm based on extended Kalman filter algorithm.In order to improve the robustness and navigation accuracy of the inertial navigation system,this thesis proposes an indoor pedestrian inertial navigation system combining an adaptive threshold zero velocity detection algorithm and an extended Kalman filter based zero velocity correction algorithm.The zero velocity detection algorithm based on adaptive thresholds is more stable and closer to the real walking distance than its fixed-threshold zero velocity detection algorithm,with an average distance error of 7.62 m at 1.1m/s,2.1m/s,and 3.1m/s speeds,and a smaller deviation in heading.(2)For the relatively large uncertainty and random bias of inertial sensors,resulting in growing navigation bias over time.A cascaded filtered pedestrian inertial navigation system is proposed to improve the navigation accuracy of an indoor pedestrian inertial navigation system combining an adaptive threshold zero velocity detection algorithm and an extended Kalman filter-based zero velocity correction algorithm,which is fused with non-recursive Bayesian filter map matching.The average error of the cascaded filtering algorithm is 0.75 m under 1.1m/s,2.1m/s,and 3.1m/s motion speed conditions,and the average distance error is1.18 m in long-distance test experiments.The experimental results show that the proposed cascaded filtering inertial pedestrian navigation algorithm has good positioning effect.(3)For complex indoor environments,a multi-sensor fusion indoor positioning and navigation system based on cascaded filtered pedestrian inertial navigation and incorporating monocular visual inertial navigation is proposed based on decision level discrimination.The multi-sensor fusion indoor positioning navigation system consists of two independent indoor positioning subsystems,which are used to discriminate the use of the indoor positioning navigation subsystem by the conditions of the actual environment.In several indoor experiments,the minimum distance error of visual-inertial guidance fusion navigation is 1.86 m.The experimental results show that the algorithm can effectively construct indoor environment information and fuse inertial guidance information for localization and navigation.