A Dynamic Enhanced Robust Cubature Kalman Filter for the State Estimation of an Unmanned Autonomous Helicopter

Abstract

This paper addresses a design and application for the problem of state estimation for an unmanned autonomous helicopter (UAH) equipped with instruments including an inertial measurement unit (IMU), a magnetometer and a global positioning system (GPS). A dynamic enhanced robust cubature Kalman filter (DERCKF) is proposed in this article. First, a robust filtering strategy is formulated to provide a strong constraint for abnormal values. Second, a new robust CKF is formulated based on the spherical cubature and Gaussian quadrature rules to estimate the probability state, without requiring calculation of the Jacobian and Hessian matrices. Then, an enhanced rule is proposed to help eliminate the accuracy degradation caused by model uncertainty disturbance when the experimental platform is operating and to improve the estimation performance of the filter. Meanwhile, by detecting the system uncertainty state, a dynamic enhanced strategy is formulated to achieve automatic adjustments for the dynamic enhanced robust rule and guarantee that the DERCKF will realize valid system state estimation at all times. Finally, numerical experimental results are presented to demonstrate the effectiveness and robustness of the DERCKF.