Research On The Modified Algorithm Of Strong Tracking For MEMS Based PNS

MEMS (Micro Electro Mechanical System) has the advantages of small size, light weight and low power consumption, which can further promote the development of PNS(Pedestrian Navigation System). Currently, the MEMS-based PNS has gradually become the research focus.The present PNS mainly relies on the MEMS/GPS (Global Positioning System)integrated system for the positioning. However, the GPS signal is likely to be locked in the environment like indoor, canyon or other sheltered places, where the system can't provide accurate and continuous positioning service for the pedestrian. In addition, other auxiliary schemes such as map matching, camera, the radio frequency signal or WiFi aiding MEMS were put forward. These methods can improve the positioning accuracy, but the system cost increases, and the external signal source is not widespread as well, thus resulting in the limitation of the application environment. Therefore, in order to realize the independent positioning and meet the different needs for pedestrian navigation, the research of autonomy positioning for pedestrian under passive signal has great significance.For the autonomous PNS, by placing the MIMU (Micro Inertial Measurement Unit)composed of MEMS sensors on the foot, the zero velocity update algorithm can be used to correct the positioning error of MIMU when the foot is still according to the pedestrian gait characteristics. The zero velocity update algorithm can effectively restrain the navigation error drift of MIMU with the velocity errors as the observation, nevertheless the unobservability of heading error will further lead to the positioning error of the pedestrian navigation system. To improve the tracking precision of the PNS and solve the problem of the heading error drift,we can make full use of the heading information provided by the magnetometer and take the magnetic heading as the reference for attitude to correct the navigation error. This paper adopts an error correction algorithm based on multiple observations, on the basis of the zero velocity detection, the information of MEMS accelerometer, gyroscope and magnetometer is fused, and the error quaternions are taken as the observation, meanwhile, the kalman filter is used to estimate the navigation parameters in order to enhance the observability of the system and improve the tracking precision of pedestrian.For the magnetic interference environment such as the indoor environment,the influence of the magnetic disturbance should be fully considered, so the magnetometer can not be taken as the reference source for heading. To improve the reliability of observation, the improved navigation algorithm based on the weak magnetic field detection is proposed for navigation and positioning of the pedestrian, under the condition of the weak magnetic field, the error of magnetic field intensity and its change rate can reflect the attitude and gyro errors of the system. Hence, we consider to use the magnetic field intensity information of the magnetometer output as the observation of the system,with the magnetic field strength error in the navigation system and its change rate in the carrier system as the observation, this model can not only provide auxiliary information for PNS, but also enhance the credibility of the observation and improve the positioning accuracy.Finally, to verify the validity of the algorithm, the field test was carried out using the MIMU of Xsens company in Dutch. Both outdoor and indoor test were taken, and we chose several different trajectory paths in order to verify the validity and applicability of the algorithm. The results showed that in the corresponding environment, the algorithm can reduce the positioning error and suppress the navigation error drift.