MEMS-SINS navigation method aided by vehicle model

In view that the frequent outages of GNSS in urban environments can quickly degrade the performance of MEMS-SINS, a new MEMS-SINS navigation method for land vehicles is proposed based on the vehicle constraints and combined with the four-channel ABS wheel speed sensors and steering angle information. By analyzing the vehicle turning and constraint characteristics, the angular velocity and acceleration are constructed as the measurements to achieve on-line compensation for MEMS's rapid drifting errors. Three-dimension vehicle-body velocity provided by ABS information and non-holonomic constraint is applied to further maintain the update of the integration Kalman filtering during GNSS outages. The road-test results demonstrate the proposed method can effectively reduce the rapid accumulation errors of SINS due to low-cost MEMS inherent bias in the circumstances of long-time outrages of GNSS. Compared with conventional body velocity constraint and odometer algorithm, the heading accuracy is improved by 70%, and the accuracies of position and velocity are also improved.