一种基于改进的“当前”统计模型自适应跟踪算法  

An Adaptive Tracking Algorithm Based on Improved "Current" Statistical Model

作　　者：钱广华[1] 李颖[1] 骆荣剑[1] 

机构地区：[1]中国人民解放军重庆通信学院,重庆400035

出　　处：《科学技术与工程》2013年第15期4191-4196,4200,共7页Science Technology and Engineering

基　　金：国家自然科学基金(61272043);重庆市自然科学基金重点项目(CSTC2011BA2016)资助

摘　　要：针对工程应用中"当前"统计模型对机动频率和最大加速度经验值依赖过大,难以根据目标的加速度变化进行实时动态调整优化的问题;以及标准"当前"统计模型在跟踪非机动或弱机动目标时,精度不高的问题,在分析机动频率物理含义及其与加速度变化关系、卡尔曼滤波的新息与加速度方差关系的基础上,提出了一种高效的机动频率和加速度方差双变量自适应算法。仿真结果表明该算法能够很好地自适应目标的加速度变化;并能有效提高跟踪精度,大大提高了对非机动或弱机动目标的跟踪精度。The standard adaptive Kalman filter algorithm based on the “current” statistical model （AF） has the problem of selecting Maneuvering frequency and maximum acceleration based on experience, and the problem of low accuracy in tracking non-maneuvering or weak maneuvering target. By analyzing the physical meaning of maneuvering frequency and relationship with acceleration, the maneuvering frequency adaptive algorithm are obtained. By analyzing the relationship between Kalman filtering innovation and acceleration variance, the acceleration variance adaptive algorithm are obtained. The contrasting simulation results of AF （Adaptive Filtering Algorithm） and MAF （Mending Adaptive Filtering algorithm） have showed MAF’s validity. MAF algorithm also obtains a better tracking accuracy, especially for non-maneuvering or weak maneuvering target, and makes “current” statistical model more conveniently for engineering application.

关 键 词：机动目标跟踪 “当前”统计模型 机动频率自适应 方差自适应 

分 类 号：TN957.52[电子电信—信号与信息处理]