Research And Implementation Of Indoor Navigation System Based On Pedestrian Dead Reckoning

In recent years,location based services have gradually penetrated into every aspect of people's daily life,bringing great convenience to people's work and life.With the increasing of indoor activities,people have put more demands on indoor positioning.However,due to the obstruction of the wall and the indoor complex environment,GPS positioning technology which is widely used in outdoor positioning is not suitable for positioning in the room.Currently there are numerous technologies that are suitable for indoor positioning,including the technology based on radio frequency identification,infrared,Bluetooth,Ultra-wideband,ZigBee,and inertial sensors and so on.In view of the popularity of smart phones,this thesis uses the pedestrian dead reckoning method to realize positioning which utilizes the sensor data in the smart phone to calculate the relative position.The main research contents and implementation work are as follows:1.Research the existing indoor positioning technology and implement an indoor positioning system based on pedestrian dead reckoning on an Android smart phone.The key technologies used in the implementation of the system include gait detection,step size estimation and heading estimation.Taking into account the differences in the user's body shape and walking posture,based on the traditional peak detection method,this paper uses the four walking characteristics periodicity,similarity,continuity,and symmetry to constrain the gait detection,eliminate pseudo-peaks,and improve step-counting accuracy.Different users are requested to take phones in different postures.And compare our method with the traditional peak detection method.2.The existing step size estimation model is studied,and the Kim algorithm is enhanced by using the continuity of the step length in the walking process of the user.So the performance of the step size estimation is improved to some extent.3.The methods which is commonly used in heading estimation are studied.In this thesis,the joint heading estimation is performed using a directional sensor and a gyroscope.The gyroscope data are converted to a navigation coordinate system by a quaternion method,because the gyro data are in the reference coordinate system.This thesis improves the original heading estimation method by weighted fusion algorithm and Kalman filter.4.Correct the position of the user by particle filter and map matching.Reduce the weight of the particle which is through the wall.And the landmark matching is a heuristic method.5.An indoor positioning system based on pedestrian dead reckoning was designed and implemented on an Android smart phone.The performance of each strategic technology and commonly used methods was tested by MATLAB.And the system is actually tested on an Android smart phone.