Uniting Attitude Estimation With Global Asymptotic Stability
Xin Tong, Junyan Yan, Shing Shin Cheng
Abstract
This paper approaches the problem of attitude estimation using biased gyro and body-referenced measurements of inertial vectors. First, a new observer for attitude estimation is proposed through introducing a family of synergistic potential functions with explicit expression of synergistic gap. Furthermore, it is convenient to adjust the design by adding additional state-dependent gains without adapting the switching rules. The proposed observer exhibits faster convergence for large attitude estimation errors than similar work in literature. Second, to achieve better steady-state performance, we unite this global observer with a traditional local observer, i.e., multiplicative extended Kalman filter. The new uniting observer inherits the desired transient and steady-state performance from the two sub-observers respectively. The performance of proposed methods has been evaluated via simulation and experiments.