Research On Human Positioning Technology Based On Combination Of Wearable Micro Inertial Sensors

Personnel positioning,also known as personal or pedestrian positioning,refers to the technique of providing spatial location information to a target under a defined reference coordinate system.In the outdoor environments,the common satellite positioning systems have localized blind spots due to loss of satellite signals under remote and occluded environmental conditions.In indoor environment,common WIFI positioning,Bluetooth positioning and other positioning technologies are severely dependent on the site environment.At present,the inertial positioning technology using a combination of micro-inertial sensors which is independent of the external environment,has strong anti-interference and stability,and can realize the acquisition of its own position information at all time.In conventional inertial positioning technology,the navigation-unit mostly worn in a single position for positioning,that limits the pedestrian biological characteristics and reduces the user's experience for the wearable device.This thesis designs the personnel positioning technology in the multiple wear modes of the positioning terminal.The main research contents are as follows:1.The micro-inertial sensor combined positioning theory is introduced and the algorithm structure and hardware structure of the inertial positioning system used in this thesis are expounded.At the same time,the basic principles and key technologies in inertial positioning are expounded,and the pedestrian trajectory estimation algorithm is deduced.2.The common error sources of inertial sensors are analyzed.And the installation error and scale factor of the three-axis MEMS accelerometer and the three-axis MEMS gyroscope are compensated by the six-position method and the least square method respectively,thereby the accuracy of the inertial sensor data is improved.By this way,the personnel location of data is more reliable.3.Combined the ergonomics characteristics and the wear mode of the MIMU,four common wear modes of MIMU are defined,that is wore in the jacket pocket,the right pocket,the right front pocket and the right rear pocket.Combined with the naive Bayesian classification principle,the three-axis MEMS accelerometer is used to combine the mean values of acceleration,root mean square and standard deviation to design the MIMU wear modes recognition algorithm based on naive Bayes.It has been verified that the designed recognition algorithm can effectively identify the MIMU wear mode and the recognition accuracy is more than 96.61%.4.For the four wear modes of MIMU,the key parameters such as gait detection,step estimation and heading estimation in the pedestrian trajectory estimation algorithm are improved.Wavelet-double threshold step counting,HDE-KF based heading optimal estimation and other algorithms are proposed.It has been verified that the algorithm can provide accurate and reliable position information in the four wear modes of MIMU,and the final comprehensive positioning error is less than 0.49%,and the compliance rate is more than 85%.