Map Aided Low-Cost MEMS Based Pedestrian Navigation Applications

Nowadays,indoor pedestrian navigation system has a big market requirement,more than 25000 developers in the world are focusing on this market.In 2012,the 12 th Five-Year Plan of the Navigation and location services technology development,which is issued by the Ministry of Science and Technology,clearly pointed out that China will promote the development of indoor positioning technology,and implement precise indoor/outdoor seamless navigation.Various kinds of techniques,such as Wi-Fi base positioning,inertial navigation based method,Bluetooth technique,map based method,vison based technique,could be used to obtain the pedestrian's position in indoor environment.To date,the major challenges for an indoor pedestrian navigation system is to reduce the cost of the system,including the time-cost and the economic-cost,without decreasing the accuracy of the system.Considering that the MEMS-based inertial sensors are small in size,light in weight,low-cost,convenient,and self-independent,and the global IMU embedded smartphone adoption rate keeping increasing year by year.Therefore,inertial navigation based method is applied in this research to obtain a primary navigation solutions.The estimated solution of inertial navigation system grows with time,and after a few minutes the system cannot work anymore.Therefore,aiding constrains and information are needed to to correct the inertial navigation errors.To take advantage of the unique kinetic characteristic features of pedestrian,Non-Holonomic Constraints and Zero Velocity updates could be used to correct the inertial navigation solution.The outputs of MEMS accelerometers are used to detect the state of motion,such as the zero velocity and the step detection.Then,these constraint information are used to correct the estimated solution of inertial navigation system through extended Kalman filter.With the development of wireless communication,the coverage rate of Wi-Fi keeps increasing.Most of the public buildings,such as universities,museums,airports,have Wi-Fi access points,and can provide free Wi-Fi signal to users.When Wi-Fi is available,the fingerprinting-based Wi-Fi position could be further used to correct the error of inertial navigation estimated solution,and solving the unobservability problem of the heading information of the inertial navigation system by using the extended Kalman filter.To solve the ambiguity problem in the Wi-Fi online positioning process,a KNN estimation method is applied to calculate the user's position and improve the estimation accuracy of Wi-Fi positioning method.Map-based positioning is a convient and low-cost method.Currently,most of the building owners and the location based service companies can provide digital indoor map information to users.Indoor map information can not only be used to present the estimated position results,but also be used to correct the navigation error during the estimation process.Furthermore,to solve the indoor map information acquisition problem,a fast and efficient floor plan processing method is presented.Particle filter can be used to solve non-linear and non-Gaussian estimation problem,moreover,it is flexible to add aiding information to the system.In this research,particle filter is applied to add indoor map information to the inertial navigation system.Different kinds of particle filter,such as traditional sequential importance particle filter,backtracking particle filter,auxiliary particle filter,are implemented,and taking the computational cost and the estimation accuracy into consideration,auxiliary particle filter is used to add the indoor map information to pedestrian navigation system.A wall cross algorithm is designed to reset the weight of particle filter,and an effective partile detection method which uses the geometrical relationship between the indoor map information and the particles is proposed.To effectively add the indoor map information to the INS,Map Matching and Map Aiding methods are novelty combined to make a full use of the free map information.To effectively combine the above information and the INS,a Kalman filter and Particle filter based cascade algorithm is proposed in this research to correct the INS error.Moreover,to take advantages of the PDR and INS method,a cascade structure algorithm is used to update the system.Real experiments with different navigation architectures and combinations are conducted in different scenarios to verify the proposed MEMS-based indoor pedestrian navigation methods.The experimental results clearly indicate that the cascade structure algorithm can decrease the computational burden of the system.Also,through the proposed methodologies,integrating indoor map information,smartphone embedded sensors,and the pre-existing Wi-Fi,the indoor position system could provide continuous,low-cost,and accurate navigation solutions for a pedestrian in indoor environments,and the RMS error of the estimated position can keep in a small value.