A Dual-Stage EKF Attitude Estimation Algorithm Based on a Six-Axis Gyroscope and Magnetometer

Le résumé fourni par la source

The inherent limitations of MEMS inertial sensors-including significant noise, bias instability, and drifting errors-severely constrain the accuracy of attitude estimation for aircraft in complex maneuvers, potentially causing traditional methods to fail. This paper addresses this critical problem through a multi-sensor fusion algorithm built upon a dual-stage Extended Kalman Filter (EKF) design. Our solution harnesses the synergistic strengths of gyroscopes, accelerometers, and magnetometers. It operates by first propagating attitude from gyroscopic rates, then executing a cascaded correction sequence: the initial stage employs accelerometer data for tilt estimation, followed by a magnetometer-based update for heading. This dualstage architecture ensures effective measurement decoupling. A key enhancement is the integration of an adaptive mechanism that performs on-the-fly optimization of the filter’s measurement-noise covariance in response to kinematic acceleration, thereby granting exceptional robustness against intense vibrations and dynamic accelerations.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.