Pedestrian Indoor Position And Navigation Based On Inertial Sensor

With the rapid development of society,the demand for precision indoor positioning and navigation technology is increasing.Due to the complex indoor environment,those methods relying on wireless network have poor stability and ineffective,such as WiFi and Zig Bee.The inertial navigation can realize the positioning and navigation without external signals.Based on IMU configured in smart mobile device,this paper studies the algorithm of indoor pedestrian dead reckoning.The main research content of this paper includes attitude error compensation.In order to overcome the attitude calculation error caused by noise accumulation,this paper determine the validity of the geomagnetic field information and the gravity acceleration information obtained by the magnetometer and the accelerometer respectively,then apply these valid data to a nonlinear Kalman filter model as observations.Thereby suppressing the cumulative error of the gyroscope and achieving a high precision attitude calculation algorithmSecondly,this paper deeply explores the general physical characteristics of pedestrians under normal walking conditions and applies them to the navigation model to improve the navigation accuracy while enhancing the robustness of the algorithm.In the step detection,this paper utilizes the characteristic that the whole body will have a small vertical displacement in the vertical direction when pedestrian is walking.This paper implements the step detection by designing the finite state machine and filters noise components by designing the filter with the normal step frequency as the parameter.Finally,this paper uses smartphone to collect inertial data and does not need to fixed inertial measurement unit.This paper abandons parallel heading assumption,and extracts the true heading of the pedestrian from the horizontal acceleration information by using the feature extraction methods commonly used in machine learning.This paper collects the inertial measurement data in a variety of behavioral modes by pedestrian handheld smart phones.The simulation results show that in the step detection algorithm,the error of detection accuracy steps is 2%,and the probability of detection false steps is far less than similar algorithms.When parallel heading assumption is not useful,the reckoning algorithm proposed in this paper is still applicable and the accuracy is between approximately 1% and 2%.