Pedestrian Indoor Positioning Based On MEMS Inertial Sensors

Using a wearable inertial/magnetic measurement unit(IMMU)for Pedestrian Dead Reckoning(PDR)has been a hot topic in commercial applications such as security applications,medical surveillance,and location based service.At present,satellite navigation systems have been widely used in military and civilian fields.However,when the receiver is obstructed by a building or other objects,the positioning accuracy of the satellite navigation system is poor or impossible to locate.The wearable IMMU-based PDR method is not affected by the blockage of buildings and therefore has been extensively studied in indoor positioning.At the same time,due to the high positioning accuracy of the positioning method based on inertial navigation in the short term,the combined positioning method of inertial navigation and other navigation methods has become a hot topic.Micro-Electro-Mechanical System(MEMS)inertial devices have the advantages of small size,low cost,and low power consumption.Therefore,the method of using MEMS inertial devices for indoor pedestrian positioning is a very practical research direction.The main positioning errors based on the MEMS inertial sensors are the error of the sensors and the error of attitude estimation.Since MEMS accelerometer have significant drift errors,the velocity obtained by integrating the acceleration has serious drift.The research contents of this thesis mainly include the following aspects: 1.Gait phase detection plays an important role in the calibration of velocity and attitude estimation.A novel gait phase detection method is proposed in this thesis.The proposed gait phase detection method has strong adaptability and it can not only provide accuracy gait estimation when pedestrian walks normally,but also when the pedestrian runs.2.In many literatures,gait phase detection and Zero Velocity Update(ZVU)methods are used to correct the drift of velocity.This thesis proposes reasonable assumptions based on the study of the drift of accelerometer.The drift of the accelerometer is calculated at the end of the swing phase,thereby the drift of the velocity during the swing phase can be calibrated.3.At present,Kalman filter and complementary filter(CF)are widely used in attitude estimation.However,the Kalman filter has a large amount of calculations,and it is complicated to model different pedestrians and different motion states.CF has a small amount of calculations and it can achieve high accuracy under low dynamic conditions.However,the performance of attitude estimation is poor under high dynamic conditions,which can not meet the requirement of accurate positioning of pedestrians in running conditions.Considering that the gyroscope has good dynamic performance,a novel attitude estimation method is proposed in this thesis based on the study of the drift of the gyroscope.The proposed attitude estimation method has the characteristics of small amount of calculations and good adaptability to different motion states.Experimental results show that the pedestrian positioning algorithm proposed in this thesis can achieve less than 2% of distance error and end-to-end position error in indoor environment,which has good practicability.