摘要
针对交互多模型(IMM)方法模型集覆盖能力与计算量相矛盾的问题,提出了将简化的Sage-Husa自适应滤波与IMM相结合,构成一种自适应交互多模型的方法。简化的Sage-Husa自适应滤波首先给出噪声统计特性的粗略值,IMM方法以该粗略值为中心,对称地得到模型集,再进行IMM估计。车载组合导航仿真表明,该算法能够以较少的模型实现对实际模态的覆盖,而且精度比IMM方法也进一步提高。
The interacting multiple model （IMM） estimator has the defect that the more sub-models, the worse real-time performance. An adaptive interacting multiple model （AIMM） algorithm is presented, which combines the IMM algorithm with the simplified Sage-Husa adaptive filtering algorithm. The Sage-Husa adaptive filter is used to estimate the rough value of measurement noise statistical characteristics. The parameters of sub-models are calculated by the rough value and then are fed into the sub-filters in IMM algorithm, where the sub-filters have different parameters of noise. Simulation results of land vehicle integrated navigation show that the AIMM algorithm can achieve the coverage of real situation through a few sub-models, and the accuracy is higher than the IMM algorithm.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第11期2070-2074,共5页
Systems Engineering and Electronics
基金
国家自然科学基金(60774065)
国家"863"高技术计划基金(2006AA12Z305)资助课题
