Research Of High-accuracy Pedestrian Navigation Algorithm Based On MEMS Sensors

With rapid development of microelectronics technology, the application area of navigation technology gradually shifted from military to the public. Indoor positioning technology is developing dramatically in recent years, attracting more and more attention from enterprises and the public. However, the disadvantage of currently existing indoor positioning system is obvious. Because most of these systems are based on infrastructure implementation, a large number of external devices should be installed in advance in navigation environment resulting in a high initial cost. In the meanwhile, the precision of these systems are waiting to be improved. In order to solve these problems, this paper will design a kind of inertial navigation system based on its sensors.Strap-down inertial navigation technology not only can avoid the defects mentioned above, but also can greatly improves the accuracy of positioning. Nevertheless, as time goes on, the system would generate an accumulated drift error. Adjusting this problem, this paper designs a kind of algorithm to effectively compensate the accumulated error based on Kalman filter. First of all, the system uses strap-down inertial calculation module to calculate the integral of inertial measurement information, meaning that the system gets attitude information from the integral of angular velocity, and transforms the coordination of acceleration through quaternion method, and gets velocity information from the integral of acceleration, and then using quadratic integral to get the position information. Secondly, the system uses the zero-velocity detection module to detect whether the pedestrians are under the zero-velocity phase through three-condition judgment method. The Kalman filter module is triggered when detects the zero-velocity phase. Finally, the system uses velocity vector calculated in the strap-down inertial calculation as measurement value and calculates the estimation of system state error by Kalman filter, then uses the covariance of the velocity estimation through forward Kalman filter to segregates the state error estimation, and further adjusts state error estimation and its covariance matrix with the backward fixed-interval smoothing technique. In the end, the system adjusts the position, velocity and attitude information of pedestrians by forward feedbacks. In addition, the author expanded Kalman filter state model by adding the zero-bias information from the accelerometer and gyroscope. This improvement further eliminated system drift error because of the feedbacks and adjustments of the inertial measurement information. After that, the system achieves the pedestrians positioning and navigation in the indoor environment.On the basis of the above error compensation algorithm, this paper designed a pedestrian navigation system, it converge to a stable state and its position accuracy is less than1meter within500meter-long road through the experimental simulation. This system achieves the positioning target of high precision.Pedestrian navigation system designed in this paper provides a simple and effective method to the indoor positioning based on sensors, which can be further integrated, improved, and applied in environments like smart home, fire rescue, shopping mall, and parking in the garage.