Vehicle-Mounted Integrated Positioning System with Vondrak Low Pass Filter and Multi-Dynamic Constraints in Urban Shaded Environment

The Global Navigation Satellite System (GNSS) alone cannot provide high-precision and continuous positioning information for vehicles. The integration of GNSS with Inertial Navigation Systems (INS) has now been very intensively developed and widely applied in high precision positioning of vehicle and provide continuous position, velocity and attitude. However, the overall performance of low-cost GNSS/MEMS IMU frequently degrades in urban shaded environments. Traditional constraints GNSS/MIMU algorithm based on zero-velocity detection can effectively increase the accuracy of the navigation system, but easily influenced by external factors to false detection. This article aims to introduce a multi-dynamic constraints model as accurate update source for EKF to improve the accuracy of navigation solutions of a vehicle during satellites signal blockages. Firstly, we present a tightly coupled strategy to integrate GPS/BDS and INS by applying extended Kalman filter with 27-states. Then, a compositive static zero-velocity detection scheme is carried out by using the Vondrak low pass filter, GNSS/INS calculated velocity and the original data of INS. Meanwhile, dynamic ZUPT constraint model is also constructed based on the motion characteristics of vehicle. An vehicle test was performed to validate the new algorithm. The results indicate that proposed muti-dynamic constraints model with Vondrak low pass filter can effectively improve the success rate of zero-velocity detection. When the satellite signal is interrupted for 120s, the root mean square of vehicle position and velocity can be dramatically reduced.