An H-infity Filter Based Research On Fusion Location Technology Of UWB And IMU

Ultra-wide Band(UWB)technology has become a popular technology for indoor positioning due to its high positioning accuracy.However,in many indoor application scenarios,UWB measurement signals are blocked by obstacles,resulting in outliers,which can not provide stable and high-precision positioning information.Inertial measurement units(IMU)are widely used as motion sensors in robotics,such as navigation and mapping tasks.Many low-cost IMU-based microelectrom-echanical systems(MEMS)are already on the market,and smartphones and similar devices are almost all equipped with low-cost embedded IMU sensors.However,the low cost IMU is affected by the system error caused by imprecise scale factor and axial error,which reduces the accuracy of position and attitude estimation.Single indoor positioning technology can not provide the positioning accuracy and stability required by practical applications.Based on this problem,this paper proposes two fusion filters based on UWB/IMU to track position,velocity and direction.In this paper,UWB and IMU data are fused based on extended Kalman filter(EKF)andH_?filter respectively.By comparing the performance of UWB,extended Kalman filter andH_?filter,the positioning accuracy of the positioning system is improved.Because UWB original data and IMU original data have some noise,this paper designs an algorithm to filter UWB and IMU data.For UWB data,zero value,mutation value and continuous invariant value are mainly filtered.Aiming at IMU data,IMU error model is established,error sources are analyzed,and calibration algorithm is designed.Calibration parameters are obtained offline by MATLAB to calibrate IMU data.In this paper,a positioning module is designed to verify the algorithm,including hardware module and software module.Hardware design includes STM32F411 main control chip peripheral circuit design,UWB(DWM1000)module,inertial navigation sensor(MPU9250),and power module.The software design is divided into two parts:label and base station firmware design and positioning system software design.The main functions of the label and base station firmware are:preprocessing and filtering the data collected by UWB,correcting the data output by IMU,and sending the preprocessed distance and attitude data to the upper computer.Since the computing power of STM32F411 chip is far inferior to that of PC,we put the final filtering and position calculation on the upper computer for processing.In addition,the upper computer can also display the real-time position of the label through the image and interact with people through the graphical interface.