Research And Implementation Of Indoor Positioning Method Based On Pedestrian Dead Reckoning

With the rapid development of information technology and the popularity of smart phones,location based services have played an important role in more and more applications.Indoor location technology has gradually become the focus of research.When it comes to the positioning method of pedestrian dead reckoning,we can just use the inertial sensor,which is low cost in hardware and easy popularization,to get the travel trajectory,through calculating the length of each step and the moving direction.However,these methods didn't consider the location change of the sensor during pedestrian walking,which led to larger estimation error.The goal of this thesis is to solve the problem that the mobile phone will swing in the pocket during movement,which led to larger positioning error.This thesis focuses on step length estimation and heading estimation algorithm in PDR system.Inertial measurement unit of smart phone is used to acquire sensor data during pedestrian walking.Pedestrians visually place their mobile phones in their trouser pockets.This thesis adopts step detection and step length estimation algorithms based on leg swinging angle.When the cell phone is placed in pedestrian's pocket,the phone inevitably inclined.It will cause more error when the data of the leg swinging angle measured by the mobile phone is directly applied to the PDR system.In this thesis,we established the mathematical model to analysis the relationship between the measured angle value and the actual leg swinging angle when the mobile phone was inclined,and proposed an improved step length estimation algorithm,which was more accurate.The mobile phone will swing when pedestrian moves on curved paths,the headi-ng estimation algorithm analyzed the turn insufficient effect,but the angle correction method based on heuristic offset elimination existed significant limitations.In this thesis,using the sensor fusion strategy,the angular acceleration in turning was corrected by the gravity acceleration sensor data,so the estimation of the turning angle was more accurate.The experimental results show that when the pedestrian walked 70 meters in the straight path,the average relative error of the step length estimation was less than 2%.When the mobile phone inclined,the average relative error of the step length estimation decreased more than 10%with implement of the improved step length estimation algorithm,and the average relative error of direction estimation decreased more than 5%with implement of the improved heading estimation algorithm.The estimation error of pedestrian step length and direction meets the requirement of design index.