| With the development of society and the progress of technological,people's demand for location and navigation information is increasing.Because of the complex and changeable of indoor environment,the propagation of electromagnetic wave is affected by attenuation,multipath and interference.The location technology based on different kinds of wireless networks can't obtain satisfactory position results.The method of positioning and navigation based on MEMS inertial sensor based pedestrian dead reckoning has the advantages of autonomous navigation,no interference of environment and no need to deploy additional infrastructure and has received extensive attention and research.Inertial sensor is widely used in positioning and navigation for its advantages of low cost and small size.The signal to noise ratio of the MEMS inertial sensor is low and the drift is large.And the cumulative error of the signal in the positioning and navigation process is increasing with the time.This has brought the challenge to the estimation of MEMS inertial sensor based pedestrian dead reckoning.Aiming at the cumulative error of signal,this paper presents angular rate moving variance detector based static detection to achieve zero-velocity update.In this paper,angular rate moving variance detector based static detection is used to detect stationary phase of human gait.Extended Kalman filter is used to realize zero velocity update,reduce the cumulative error of signal and improve positioning accuracy.Aiming at the low signal noise ratio of the MEMS inertial sensor and large drift,this paper presents Shannon entropy based EMD algorithm to denoise for MEMS gyroscope signal.Though there is large non-linear and non-stationary drift noise in MEMS gyroscope signal,Shannon entropy based EMD algorithm can denoise the gyroscope signal effectively and improve the accuracy of pedestrian trajectory.Based on the MATLAB platform,the algorithm of indoor pedestrian trajectory estimation is realized.Angular rate moving variance detector based static detection is used to achieve zero-velocity update and Shannon entropy based EMD algorithm is used to denoise.The data collected through MTx inertial sensor is used to verify.The results show that the method proposed in this paper can improve the accuracy of pedestrian trajectory estimation effectively. |
