噪声统计特性失配的鲁棒目标跟踪方法  

Robust target tracking method for noise statistical characteristics mismatch

作　　者：周飞菲[1] 

机构地区：[1]郑州升达经贸管理学院信息工程学院

出　　处：《电子测量与仪器学报》2018年第9期195-201,共7页Journal of Electronic Measurement and Instrumentation

基　　金：河南省科技厅科技攻关项目(豫科鉴委字[2013]第521号);郑州升达经贸管理学院创办人科研基金项目(SDXM2015050)资助

摘　　要：针对噪声统计特性失配情况下目标鲁棒跟踪问题,提出了一种噪声失统计特性失配的新型卡尔曼滤波器跟踪方法。该方法首先采用渐消因子对状态估计一步预测的均方误差进行修正,通过新息残差矢量对渐消因子进行在线自适应更新,克服噪声统计特性失配引起的模型漂移问题;接着,利用渐消记忆指数加权的方法实时估计新息残差矢量,克服传统利用加权求和方法计算新息残差估计的精度较差的问题,进一步提升新息残差的估计精度。实验结果表明,在噪声失配的情况下,新方法的性能优于传统的卡尔曼滤波器和渐消因子卡尔曼滤波器,具有较高的跟踪精度和鲁棒性,是一种优秀的跟踪方法。In order to solve the problem of target robust tracking problem under the condition of noise statistical characteristic mismatch,a new Kalman filter tracking method based on noise statistics characteristics mismatch is proposed. First,the fading factor is used to correct the mean square error of one step prediction of the state estimation,and to update the fading factor online by updating the residual vector,which overcomes the model drift problem that caused by the statistical characteristic mismatch of the noise. Second,the fading memory index weighting method is used to estimate the new residual vector in real-time,and to overcome the poor accuracy of the traditional weighted residual method,which further improves the estimation accuracy of the new residuals. The experimental results show that the performance of the new method is better than that of the traditional Kalman filter and the fading factor Kalman filter. It has high tracking accuracy and robustness and is an excellent tracking method.

关 键 词：目标跟踪 噪声统计特性 卡尔曼滤波器 新息残差 记忆因子 

分 类 号：TN713[电子电信—电路与系统]