基于消费级惯性传感器的行人导航系统研究

Research on Pedestrian Navigation System Based on MEMSInertial SensorAbstractWith the rapid development of microelectronics technology,the popularity of locationservice has brought great convenience to people's life.In the outdoor environment,userscan obtain real-time,high-precision positioning technology services by dragging thecomplete GNSS.Indoor positioning technology is generally used in environments whereGNSS cannot be applied,such as buildings,tunnels,forests or between buildings.However,the indoor positioning system using active signal needs to install the infrastructure inadvance,which is often helpless when confronted with signal interference and someemergency situations.Based on this situation,this paper adopts an inertial sensor that canrealize autonomous navigation by data acquisition,and explores the use of MEMS inertialsensor measurement unit for pedestrian indoor navigation.Based on the heading reckoningof foot bound pedestrian,the positioning accuracy and system stability of foot boundpedestrian navigation system are improved by improving data quality,improving theaccuracy of gait detection and navigation information fusion as the research objectives.Themain research work of this paper is as follows:Firstly,on the basis of not assuming the noise distribution model,the constrained totalleast squares algorithm and the iterative least squares algorithm are applied to realize thereal-time calibration of the three-axis accelerometer,which improves the real-time andautonomy of the calibration.It is more stable in practical application.Secondly,the 9-axis data (acceleration,angular velocity,magnetic data)are measuredusing the consumer grade inertial sensing chip MPU9250,and the data are sent to thenavigation computer through the serial port.Aiming at the gait detection module inpedestrian navigation algorithm,a method of pedestrian gait detection using gyro dataalone is proposed.The gyro noise and bias error of consumer grade inertial sensors affectthe heading divergence.The traditional Zupt can suppress the velocity error and horizontalattitude Angle to a certain extent,but there is still no proper correction method for theheading Angle.To solve this problem,a heading self-observation method is proposed.The