基于模糊神经网络融合的目标自适应跟踪算法  

Adaptive algorithm in target-tracking based on fuzzy neural networks information fusion

作　　者：张圆[1] 郭晨[1] 谢春思[2] 梁勇[2] 

机构地区：[1]大连海事大学自动化与电气工程学院,辽宁大连116026 [2]海军大连舰艇学院装备系统与自动化系,辽宁大连116018

出　　处：《大连海事大学学报》2007年第2期120-123,共4页Journal of Dalian Maritime University

基　　金：国家自然科学基金资助项目(60474014);高等学校博士学科点专项科研基金资助项目(20040151007);交通部应用基础研究资助项目(200432922504)

摘　　要：为解决基于“当前”统计模型的自适应滤波器对弱机动,特别是非机动目标跟踪精度下降问题,提出基于模糊神经网络的目标自适应跟踪算法,对并行工作的两滤波器进行数据融合.仿真结果表明:与一般自适应算法相比,该算法对各种机动程度的目标跟踪精度均有不同程度的提高,能更好地适应目标的各种运动形式,尤其适用于对目标的速度和加速度估计精度要求较高的场合,在指控、火控系统中具有实用价值.This paper presented a target adaptive tracking algorithm based on fuzzy neural networks to solve problem of tracking precision drop for low-maneuvering and non-maneuvering targets based on current statistic model. The method made a data fusion of two filters working in parallel. Simulation results show that the proposed algorithm can improve target-tracking accuracy for different kinds of maneuvering degree and adapt to all kinds of movement forms of the target, especially to the situation when high filtering precision for velocity and acceleration of the target is needed. It has practical value of reference for the research of fire control system and command control system.

关 键 词：机动目标跟踪 模糊神经网络 自适应算法 卡尔曼滤波 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]