Fast identification of a detailed two-track model with onboard sensors and GPS

Abstract

This contribution investigates the sensor fusion of a three antenna GPS-system (position, velocity, yaw, pitch and roll angles) and an inertial measurement unit (IMU, angular velocities and accelerations) to get a three dimensional motion estimation of a vehicle. To handle the different delays of GPS measurements and the nonlinearities in the measured system a Sequential Enhanced Kalman Filter is used. Bias and gain errors of the IMU are estimated to allow a precise calibration of angular velocities and accelerations.