Vehicle State Estimation Based on the Integration of Low-cost GPS and INS

摘要

GPS is now widely used in vehicle navigation.Compared to the location information,the velocity information from GPS is rather accurate even for a low-cost one.The low-cost GPS usually has a long sampling interval which will cause the accumulation of INS bias.INS bias can be divided into random bias (high-frequency) and steady bias (low-frequency).This article proposes a vehicle state estimation method hased on the integration of low-cost GPS and INS.An adaptive low-pass filter is established to remove the effect of random bias.Using the accurate velocity information from GPS,the steady bias of INS can be estimated through two Kalman filters.The longitudinal/lateral velocity and vehicle sideslip angle can be obtained by the integral of corrected INS signals with GPS signals as initial value.The proposed estimation method is verified through simulation in different conditions.The simulation result shows that the RMS error of vehicle sideslip angle estimation is less than 0.51°.