CAJ | 학술논문

针对 SINS／GPS 直接式组合导航姿态估计中，GPS 信号易受外界干扰而引入污染观测量等问题，提出基于 Huber 的鲁棒化四元数无味卡尔曼滤波算法。通过研究在四元数无味卡尔曼滤波算法中引入 Huber 鲁棒化框架，对原算法的量测更新进行修正，增强滤波算法的鲁棒性与稳定性。以 SINS／GPS 位置松组合为应用背景，在不同仿真环境下，验证提出算法对姿态信息的估计效果。试验结果表明：与原算法相比，该算法具有更好的鲁棒性和稳定性。
침대 SINS／GPS 직접식조합도항자태고계중，GPS 신호역수외계간우이인입오염관측량등문제，제출기우 Huber 적로봉화사원수무미잡이만려파산법。통과연구재사원수무미잡이만려파산법중인입 Huber 로봉화광가，대원산법적량측경신진행수정，증강려파산법적로봉성여은정성。이 SINS／GPS 위치송조합위응용배경，재불동방진배경하，험증제출산법대자태신식적고계효과。시험결과표명：여원산법상비，해산법구유경호적로봉성화은정성。
Considering that GPS observations can be easily disturbed in the attitude estimation of inte-grated SINS/GPS,this paper proposes a novel Huber-based unscented quaternion filter.The Huber technique is used to robustify the unscented quaternion filter for modifying and updating the original measurement so as to improve the robustness and stability of the filter.The traditional filter and pro-posed filter are evaluated under different simulating conditions for the attitude estimation of the inte-grated SINS/GPS loosely-coupled system.The results prove the proposed filtering algorithm to be su-perior to the traditional filtering algorithm in terms of robustness and stability.