AUV平台的水下机动目标跟踪  

Underwater Maneuveing Target Tracking on AUV Platform

作　　者：陈立军[1] 郝保雷 郑剑 CHEN Lijun;HAO Baolei;ZHENG Jian(No 92682,Chinese Liberation Army,Zhanjiang,524000,China)

机构地区：[1]中国人民解放军92682部队,湛江524000

出　　处：《网络新媒体技术》2020年第5期48-54,共7页Network New Media Technology

摘　　要：提出了一种过程噪声自适应调整的目标跟踪算法。利用水下航行器配备的探测声纳,根据目标径向速度、距离以及方位信息,结合航行器自身的导航信息,对目标进行高精度跟踪。传统交互多模型算法的主要缺点在于模型切换速度较慢,导致对目标机动的检测有明显滞后,影响跟踪精度。通过引入不同的系统噪声模型,对机动目标的不同状态进行建模,即目标运动模式切换时自动采用高过程噪声模型,运动模式保持时采用低过程噪声模型。这样,在不降低系统跟踪精度的情况下,有效提高了系统对机动目标的跟踪速度。仿真表明,本算法明显优于传统交互多模型算法。An underwater maneuvering target tracking algorithm is proposed for Autonomous Underwater Vchicle(autonomous undorwator uchocle,AUV)platforms equipped with sonar systems.Radial velocity is provided,together with distance,bearing and navigation information to obtain high precision target tracking.Meanwhile,an improved Interactive Multiple Model(IMM)filter with different levels of process covariances is introduced to deal with target maneuvering problem,using the fact that process covariance in the filter is a trade-off between response speed and tracking error.It is shown that the improved IMM filter can automatically switch to the model with higher process covariance when the maneuver occurs,which can improve the tracking speed;and switch to the model with lower process covariance when its maneuver is finished,which means we can obtain higher accuracy in target tracking.Simulation results illustrate that the improved IMM has a better tracking performance compared to conventional IMMs.

关 键 词：目标跟踪 交互多模型 无迹卡尔曼滤波 

分 类 号：U674.941[交通运输工程—船舶及航道工程]