Research On Multi-sensor Information Fusion And Attitude Measurement For Micro Inertial Components In Embedded Inertial Navigation System

In this paper,high-speed rotating projectile,small UAV and other micro-inertial navigation fields for the research background,to carry out small-volume micro-inertial components attitude measurement research.At present,there are many ways of attitude measurement.The single sensor attitude measurement method can not meet the requirement of attitude measurement accuracy.In order to acquire the attitude information more accurately,the attitude measurement method based on multi-sensor information has become a hot but difficult issue in current research.Based on the micro-inertial measurement components in embedded navigation system,the research of multi-sensor information fusion attitude measurement method is carried out.A single sensor attitude measurement algorithm using gyroscopes,accelerometers and magnetoresistive sensors in a micro inertial measurement unit is completed.The emphasis is put on how to estimate the state quantity of Kalman filter,and the method to estimate the attitude angle by using the fused gyroscope,accelerometer,magnetoresistive sensor and satellite velocity information is realized.The state space model for estimating attitude angle under high dynamic environment is established by satellite velocity information.The mean square error of system process noise and observation noise is adjusted by fading memory weighting method.The statistical characteristics of noise in filtering process are solved without transcendental knowledge and prophetic question.According to the output signals of gyroscope and magnetoresistive sensor collected online,a one-step prediction error is estimated based on RBF neural network to ensure the filtering accuracy of the filter during the satellite signal lock-out time.Independently developed embedded navigation system including navigation computer and micro inertial component,built embedded hardware and software architecture based on multi-sensor information fusion,and realized the acquisition and processing of multi-sensor data,the reception and analysis of satellite data,attitude prediction,filter processing and other tasks.The vehicle attitude measurement experiment is carried out.The experimental results verify the effectiveness of the improved dynamic attitude measurement method.