基于改进CS模型的自适应概率数据关联算法  

Adaptive Probability Data Association Algorithm Based on Modified "Current" Statistical Model

作　　者：戴路[1] 黄双华[1] 谢永亮[1] 

机构地区：[1]海军工程大学电子工程学院,武汉430033

出　　处：《电光与控制》2010年第8期37-40,共4页Electronics Optics & Control

基　　金：国防预研基金(4010501020103)

摘　　要：针对密集杂波环境下传统概率数据关联算法对突发机动目标跟踪性能下降问题,提出了一种基于采用渐消因子的改进"当前"统计模型的自适应概率数据关联算法。该算法改进了传统的"当前"统计模型中加速度方差的计算方式,并在滤波算法中采用了渐消因子,克服了传统卡尔曼滤波的3大缺陷,通过改变预测协方差来修正滤波增益,在保持跟踪精度的前提下,能自适应调整滤波器带宽,增强了系统对突发机动的跟踪能力。理论分析和仿真结果表明,该算法比采用强跟踪滤波器的概率数据关联算法更有效。To improve the performance for tracking a sudden maneuvering target in cluttered environment,a new adaptive Probability Data Association（PDA） algorithm is presented based on the modified＂Current＂ Statistical（CS） model that uses a fading factor.By modifying the computing method of traditional CS model＇s acceleration variance and using a fading factor in filtering,the defects of traditional Kalman filtering were restrained.The filter gain was modified by changing forecast covariance,thus the filter bandwidth could be adjusted adaptively while keeping the tracking precision.Therefore the performance for tracking strong maneuvering targets was improved.Theoretic analysis and simulation results showed that this algorithm is more effective than PDA based on strong tracking filter.

关 键 词：目标跟踪 当前统计模型 强跟踪滤波器 自适应跟踪 概率数据关联 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TN953[电子电信—信号与信息处理]