Research On Large-scale MIMU Array And Its Data Fusion

With the rapid development of science and technology,people's demand for indoor positioning is increasing.Due to its good performance in autonomy and safety,inertial positioning has become a popular solution for indoor positioning.However,the inertial positioning requires the integration of IMU(Inertial Measurement Unit),which makes the positioning data of the inertial navigation system diverge,resulting in poor long-term accuracy and difficult to be widely used.The poor long-term accuracy of the inertial positioning is caused by the drift of the gyroscope of the IMU.High precision IMU is cost highly and it is difficult to apply to personal positioning.This paper focus on the MIMU(Micro Inertial Measurement Unit)array and virtual gyro data fusion algorithm explore a high-quality data processing method for MIMU array,so as to provide a low-cost and high-precision solution for personal indoor positioning.The main works of this paper are as follow:(1)A large scale MIMU array is designed and constructed and its embedded software is developed.The large-scale MIMU array is composed of control system and MIMU array,among them,the control system is STM32 embedded system as the core control device,and the MIMU array adopts 32 MEMS inertial sensor schemes to constrain the noise of its gyroscopes by means of redundant measurement.In the large-scale MIMU array hardware system,the large-scale MIMU array embedded software is developed,the system control and MIMU array data acquisition function is completed,which has the advantages of simple control,stable sampling rate.(2)In order to improve the gyro precision of low-cost MIMU,a large-scale MIMU array data fusion algorithm based on Kalman filter is proposed,and a gyro fault detection algorithm is added.The gyro error model is established and a virtual gyro is obtained by Kalman filter fusion,and analyzes its experimental results.When the algorithm is applied to the large-scale MIMU array,the zero-bias instability of the virtual gyro is reduced to 1/6.08 of the average value of a single gyro in the static experiments.The 1? standard deviation is reduced to 1/4.33 of the average value of a single gyro in dynamic experiments.(3)In order to verify the effectiveness of the data fusion algorithm of large-scale MIMU array,experiments such as error calibration,fusion effect and fault detection were carried out and the experimental results were analyzed.Firstly,the zero bias,zero bias stability and Allan variance curves of the gyro of MIMU array are tested through an experiment.Secondly,the alignment error and scale factor error of large-scale MIMU array are calibrated experimentally and the calculation process is simplified.Then,on the basis of calibration and compensation of gyro array errors,the static and dynamic experimental evaluation of the data fusion algorithm of virtual gyro is carried out.In this paper,a large-scale MIMU array is constructed to provide a low-cost scheme for personal indoor positioning,and a data processing scheme for large-scale MIMU array is proposed,whose effectiveness and feasibility are verified by experiments.It provides a complete solution for improving the gyro precision of low cost MIMU.