Study On Pedestrian Positioning Technology Based On MEMS Inertial Sensors

Global navigation satellite system(GNSS)is widely used in the field of vehicle navigation and aerospace.However,in indoor or unban tunnel,the satellite signal can be severely weakened due to the shielding.Thus,the problem of positioning has been increasingly emerging in GNSS-dark ares.As it is dependent of installation of infrastructures and any external information,the self-contained positioning technology based on MEMS inertial sensors has become a research hotspot in the field of indoor positioning.Aiming at the current critical problems such as positioning error accumulation and gyroscope heading drift of indoor positioning system based on MEMS inertial sensors,this paper focuses on research of the adaptive zero-velocity detection method,gyroscope heading drift correction algorithm and magnetometer calibration method.The main content and achievements of this paper are summarized as follows:1.The adaptive zero-velocity detection algorithm is studied.Existing zero velocity detection methods using fixed threshold cannot provide good performance at high gait speeds or stair climbing.An adaptive zero-velocity detection algorithm based on multisensor fusion is proposed.Firstly,the measurements of accelerometer,gyroscope and pressure sensor are employed to construct an optimal zero velocity detector based on the Neyman Pearson criterion with predetermined false alarm probability.Secondly,by analyzing the relationship between the motion mode and the angular rate,a method of adaptive detection threshold is proposed and the relation function between the angular rate and the optimal detection threshold is established.Finally,to solve the problem of height drift during stair movement,a method of stair movement recognition is proposed and an adaptive threshold function under stair movement mode is established.The experimental results demonstrate that the proposed is effective and efficient.2.The headind correction method of dual foot-mounted positioning system is studied.Existing headind correction methods of dual foot-mounted positioning system are based on the maximum or minimum range constraint between the two feet.These method cannot correct the heading error if the characteristics of gyro drift of dual foot-mounted positioning systems are the same.In order to solve this problem,a heading correction method of dual foot-mounted positioning system based on building prior knowledge is proposed.Firstly,a heading correction method of single foot-mounted positioning system is proposed based on the structural characteristics of buildings.The straight-motion state is detected according to the proposed constraints,then the heading error of the positioning system is corrected by using the building reference direction.On this basis,the heading correction method of the dual foot-mounted positioning system based on the range constraint and the heading error difference constraint is put forward to correct the headinf error when the motion state does not satisfy the straight constraint.3.The calibration method of magnetometer is studied.Existing magnetometer calibration methods are all executed in free space.However,the positioning system studied in this subject is installed in the shoes.Hence,if the magnetometer is calibrated in free space and then installed in the shoe,it will suffer from additional interference,which cannot be compensated by the traditional magnetometer calibration method.Aiming at this problem,this paper proposes an online calibration method of magnetometer.By analyzing human motion model,the calibration problem of magnetometer is transformed from three-axis calibration to two-axis calibration.The proposed method can eliminate the interference of the shoe to the magnetometer effectively and simplify the calibration procedure significantly.4.The headind correction method based on quasi-static magnetic field detection is studied.In indoor environment,the magnetic field suffers from severe interference.Using the disturbed magnetic information may induce extra errors during the estimation of the gyroscope's errors.To address this problem,a method of quasi-static magnetic field of detection is proposed in this paper.A detector of quasi-static magnetic field is deduced by combining the change rate of magnetic field amplitude,geomagnetic inclination,the difference between magnetically derived heading and gyroscope heading.The quasi-static magnetic field detected by the proposed detector is used to calculate the heading error which is fed into the zero velocity updated-aided extended kalman filter to correct the heading drift of gyro.