Research On Inertial Navigation Aided By Magnetometer And Land Vehicle Dynamics

In view of the problem that the stand-alone inertial navigation error of the land vehicle will diverge rapidly when the GPS fails,this paper studies a new vehicle navigation algorithm based on the magnetic and vehicle dynamics.This method includes proposing a land vehicle dynamics identification method based on support vector machine,using only the output of inertial measurement unit.This method extracts and optimizes 3 time domain and frequency domain features to help identify the stationary and straight-line motions in real-time,and the cornering state is determined by comparing the outputs of Z-axis gyroscope and the threshold.Then,different observations utilizing the outputs of inertial measurement unit based on the stationary,straight-line and cornering motions of land vehicles are calculated.These observations are used as redundant information of a subsystem.Combined with the output of magnetometer and error model of strap-down inertial navigation system,a Kalman Filter is designed to estimate the error of navigation parameters and then correct these navigation parameters and the effectiveness of the Kalman Filter is analyzed through simulation.To verify the validity of the whole algorithm through experiments,an experimental hardware platform of land vehicle navigation system is built.The angular velocity and angular position experiments are designed based on a three axis turntable to completely calibrate the inertial measurement unit,and a simple two step calibration method which doesn't need any auxiliary equipments is also designed to completely calibrate the magnetometer.Finally,an outdoor field test is designed to verify the effectiveness of vehicle dynamics identification method and vehicle dynamics aided land vehicle navigation algorithm.The results show that the dynamics identification method can identify the motion states of land vehicles at the frequency 1 Hz with the accuracy reaching up to 98%.With the identified vehicle motion states,the divergence speed of the inertial navigation error during the GPS failure can be restrained,and the specific performance is related to the motion states of the vehicle.When the vehicle is at stationary or cornering motions,the inhibitory effect is better,otherwise,the results are slightly worse,but compared to the vehicle pure inertial navigation method,the accuracy of navigation is greatly improved.