期刊
APPLIED SCIENCES-BASEL
卷 7, 期 10, 页码 -出版社
MDPI
DOI: 10.3390/app7101003
关键词
maneuvering target tracking; interacting multiple model (IMM); unscented Kalman filter (UKF); Gaussian distribution; sensor fault; cubature principle; adaptive matrix gene
类别
资金
- National Natural Science Foundation of China [61601505]
- National Aviation Science Foundation of China [20155196022]
- Shaanxi Natural Science Foundation of China [2016JQ6050]
Aimed at solving the problem of decreased filtering precision while maneuvering target tracking caused by non-Gaussian distribution and sensor faults, we developed an efficient interacting multiple model-unscented Kalman filter (IMM-UKF) algorithm. By dividing the IMM-UKF into two links, the algorithm introduces the cubature principle to approximate the probability density of the random variable, after the interaction, by considering the external link of IMM-UKF, which constitutes the cubature-principle-assisted IMM method (CPIMM) for solving the non-Gaussian problem, and leads to an adaptive matrix to balance the contribution of the state. The algorithm provides filtering solutions by considering the internal link of IMM-UKF, which is called a new adaptive UKF algorithm (NAUKF) to address sensor faults. The proposed CPIMM-NAUKF is evaluated in a numerical simulation and two practical experiments including one navigation experiment and one maneuvering target tracking experiment. The simulation and experiment results show that the proposed CPIMM-NAUKF has greater filtering precision and faster convergence than the existing IMM-UKF. The proposed algorithm achieves a very good tracking performance, and will be effective and applicable in the field of maneuvering target tracking.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据