Attitude and Gyro bias estimation using range-difference and IMU measurements

Abstract

This paper considers the problem of constructing an estimator for attitude and gyro bias with close-to-optimal noise properties and proven stability. The estimator is based on measuring difference in range from three or more known, fixed positions to two or more points on the vehicle. In addition, IMU measurements are also used, and a position estimate is needed. The vectors between points on the vehicle are known in the body frame and estimated in the inertial frame using the available measurements. A non-linear observer is used to create a linearization point for a Linearized Kalman Filter taking the full, non-linear system into account. Simulations suggest that the estimator has similar noise properties as an optimal, non-implementable linearized Kalman Filter using the exact state as linearization point.