模糊化模型概率的IMM-SUPF机动面目标跟踪
发布时间：2018-11-12 16:56
【摘要】：为了提高跟踪系统对水面机动目标的跟踪能力,本文将水面目标建模为椭圆形面目标,提出一种模糊化模型概率的交互多模型(interacting multiple model,IMM)强无迹粒子滤波算法。首先,利用现代高分辨率雷达获得的面目标扩展测量,给出了基于面目标的跟踪测量方程。其次,将强无迹粒子滤波(strong unscented particle filter,SUPF)算法引入到IMM中得到IMM-SUPF。该SUPF算法利用强跟踪无迹卡尔曼滤波(strong tracking unscented Kalman filter,STUKF)产生粒子建议分布。由于STUKF采用渐消因子调整UKF的状态模型协方差和观测模型协方差的比例,使得建议分布更符合真实状态的后验概率分布,从而提高了IMM算法中子模型滤波器的估计精度。最后,基于模糊隶属度函数对粒子的模型概率进行模糊化,从而在提高真实模型滤波器中粒子模型概率的同时,减小非匹配模型滤波器中粒子模型概率,进而提高IMM算法的估计融合精度。Monte-Carlo仿真实验表明,相比于传统的基于质点目标的IMM-UPF算法,文中所提的基于面目标的IMM算法跟踪精度更高,且所提算法的误差超调量更小,收敛更快。此外,所提面目标IMM算法的跟踪精度也要高于面目标IMM-UPF算法。不同于传统的质点目标IMM算法,文中将水面目标建模为椭圆形面目标,并利用面目标扩展测量信息设计了模糊化模型概率的IMM-SUPF算法。该算法进一步提高了跟踪系统对水面机动目标的跟踪能力。
[Abstract]:In order to improve the tracking ability of the tracking system, the surface target is modeled as an elliptical surface target in this paper, and an interactive multi-model (interacting multiple model,IMM (strong unscented particle filter) algorithm with fuzzy model probability is proposed. Firstly, using the extended surface target measurement obtained by modern high resolution radar, the tracking measurement equation based on surface target is given. Secondly, the strong unscented particle filter (strong unscented particle filter,SUPF) algorithm is introduced into IMM to get IMM-SUPF.. The SUPF algorithm uses strong tracking unscented Kalman filter (strong tracking unscented Kalman filter,STUKF) to generate particle recommendation distribution. Because STUKF adopts fading factor to adjust the ratio of state model covariance and observation model covariance of UKF, the proposed distribution is more in line with the posteriori probability distribution of the real state, thus improving the estimation accuracy of the neutron model filter of the IMM algorithm. Finally, based on the fuzzy membership function, the probability of particle model is fuzzied, so as to increase the probability of particle model in the real model filter and reduce the probability of particle model in the non-matching model filter. Monte-Carlo simulation results show that compared with the traditional IMM-UPF algorithm based on particle target, the proposed IMM algorithm has higher tracking accuracy than the traditional IMM-UPF algorithm based on particle target. The error overshoot of the proposed algorithm is smaller and the convergence is faster. In addition, the tracking accuracy of the proposed IMM algorithm is higher than that of the IMM-UPF algorithm. Different from the traditional particle target IMM algorithm, the surface target is modeled as an elliptical surface target in this paper, and the IMM-SUPF algorithm of fuzzy model probability is designed by using the surface target extension measurement information. The algorithm further improves the tracking ability of the tracking system for maneuvering targets on water surface.
【作者单位】： 南京理工大学自动化学院;
【基金】：国家自然科学基金资助项目(61273076)
【分类号】：TN953
本文编号：2327661
[Abstract]:In order to improve the tracking ability of the tracking system, the surface target is modeled as an elliptical surface target in this paper, and an interactive multi-model (interacting multiple model,IMM (strong unscented particle filter) algorithm with fuzzy model probability is proposed. Firstly, using the extended surface target measurement obtained by modern high resolution radar, the tracking measurement equation based on surface target is given. Secondly, the strong unscented particle filter (strong unscented particle filter,SUPF) algorithm is introduced into IMM to get IMM-SUPF.. The SUPF algorithm uses strong tracking unscented Kalman filter (strong tracking unscented Kalman filter,STUKF) to generate particle recommendation distribution. Because STUKF adopts fading factor to adjust the ratio of state model covariance and observation model covariance of UKF, the proposed distribution is more in line with the posteriori probability distribution of the real state, thus improving the estimation accuracy of the neutron model filter of the IMM algorithm. Finally, based on the fuzzy membership function, the probability of particle model is fuzzied, so as to increase the probability of particle model in the real model filter and reduce the probability of particle model in the non-matching model filter. Monte-Carlo simulation results show that compared with the traditional IMM-UPF algorithm based on particle target, the proposed IMM algorithm has higher tracking accuracy than the traditional IMM-UPF algorithm based on particle target. The error overshoot of the proposed algorithm is smaller and the convergence is faster. In addition, the tracking accuracy of the proposed IMM algorithm is higher than that of the IMM-UPF algorithm. Different from the traditional particle target IMM algorithm, the surface target is modeled as an elliptical surface target in this paper, and the IMM-SUPF algorithm of fuzzy model probability is designed by using the surface target extension measurement information. The algorithm further improves the tracking ability of the tracking system for maneuvering targets on water surface.
【作者单位】： 南京理工大学自动化学院;
【基金】：国家自然科学基金资助项目(61273076)
【分类号】：TN953
【相似文献】
相关期刊论文 前1条
1 万锋,赵宇鹏,杨汝良;基于椭圆轨道的星载SAR面目标原始数据模拟[J];遥感技术与应用;2003年02期
，本文编号：2327661
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2327661.html
