Adaptive Kalman filter for the navigation system with virtual velocity measurement

In this paper, we propose an adaptive Kalman filter for the INS/GPS navigation system with the virtual velocity measurement under nonholonomic constraints. The constraints that govern the motion of a land vehicle are used to obtain velocity measurement during periods of GPS signal outages. The velocity measurement is contaminated by time-varying noise because the pure navigation velocity output is influenced by the sensor errors, the vehicle velocity, etc. The proposed filter can adaptively change the variance of the measurement noise using the residual of the filter and guarantee the bound of the variance. The proposed method is tested for a land vehicle using two GPS receivers and a tactical grade INS. The experimental results show that the adaptive filter provides a reasonable solution to the problem of GPS signal outage.