摘要
针对组合导航系统中出现的线性非线性混合滤波模型,提出一种新的混合高斯粒子滤波算法(MGPF).该滤波算法在状态更新过程中借鉴线性卡尔曼滤波思想直接更新状态量的高斯分布参数,而非逐个更新每个粒子,因此很大程度上减少了高斯粒子滤波算法(GPF)的计算量,同时滤波精度也有一定的提高.建立了捷联惯性导航系统与全球卫星定位系统(SINS/GPS)相结合的深组合滤波模型,并对新算法MGPF进行了仿真验证,所得结果表明了该算法的有效性.
For mixture linear and nonlinear model in integrated navigation system, a new algorithm of mixture Gaussian particle filtering（MGPF） is proposed. The stage of GPF state updating can be improved with the thought of Kalman filter （KF）. The updating stage is to update Gaussian distribution parameters of the particle rather than update all particles one by one. Compared with the traditional GPF, the novel algorithm can improve filtering precision and reduce filtering time. The MGPF algorithm is applied to SINS/GPS integrated navigation model. The simulation experiment on the established model shows the effectiveness of the algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2011年第1期85-88,95,共5页
Control and Decision
基金
国家自然科学基金项目(60702003)
航空科学基金项目(20090852012)
