Angular Velocity Estimation Using Rate-Integrating Gyro Measurements

Abstract

Rate integrating gyroscopes (RIGs) measure integrated angular rates or angular displacement, requiring an observer or filter to provide full state feedback to the attitude controller. A nonlinear observer is presented, which estimates the angular velocity of a rotating rigid body using continuous-time measurements from an RIG. Torque applied to the system is assumed to be accurately known. When the inertia matrix is accurately known, the provided observer achieves exponential convergence, while an asymptotically converging adaptive observer is designed for cases when the inertia is unknown. The nonadaptive observer is shown to be robust to bounded inaccuracies in the knowledge of inertia and external torque acting upon the system. For the adaptive observer, conventional certainty equivalence methods require the angular rate states that are unavailable. To overcome this problem a novel update rule with additional control knobs is proposed, which uses just the attitude states for inertia parameter adaptation. The observer is tested in simulation to demonstrate its effectiveness.