引用本文:杨海燕,尤政,王琳.基于传感器多模式调度的智能目标跟踪算法[J].控制理论与应用,2012,29(9):1186~1192.[点击复制]
YANG Hai-yan,YOU Zheng,WANG Lin.Smart target tracking algorithm based on multi-mode sensor scheduling[J].Control Theory and Technology,2012,29(9):1186~1192.[点击复制]
基于传感器多模式调度的智能目标跟踪算法
Smart target tracking algorithm based on multi-mode sensor scheduling
摘要点击 1486  全文点击 1262  投稿时间：2011-04-12  修订日期：2012-02-26
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  
  2012,29(9):1186-1192
中文关键词  智能目标  目标跟踪  传感器多模式调度  交互多模型跟踪算法
英文关键词  smart target  target tracking  multi-mode sensor scheduling  interacting multiple models tracking algorithm
基金项目  国家高新技术发展计划资助项目(2010AA8090514C); 航空科学基金资助项目(20095196012).
作者单位E-mail
杨海燕* 清华大学 精密仪器与机械学系
空军工程大学 工程学院 		yanghy07@mails.tsinghua.edu.cn 
尤政 清华大学 精密仪器与机械学系  
王琳 总参陆航研究所  
中文摘要
      智能目标具有反跟踪的特性, 其运动状态的改变与跟踪传感器的工作模式相关. 为了实现对智能目标的有效跟踪, 文中提出一种基于传感器多模式调度的交互多模型跟踪算法. 首先, 为智能目标构建混合多模型集, 描述其与传感器模式相关的智能运动特性; 其次, 提出一种传感器多模式调度策略, 综合考虑了一步预测的目标跟踪误差、二步预测的目标运动趋势误差以及传感器模式的辐射代价3方面因素, 据此构建传感器多模式调度的目标函数; 最后, 将传感器多模式调度与交互多模型跟踪算法相结合, 通过“模式调度!交互跟踪!模式调度”的闭环结构, 提高了对智能目标跟踪的自适应性. 仿真结果表明, 与选择固定传感器模式的方法和基于近视策略的传感器模式调度方法相比, 文中方法实现了对智能目标的持续精确跟踪.
英文摘要
      The smart target has anti-tracking characteristics. If the operating mode of the sensor for tracking the smart target is changed, the smart target modifies its motion state accordingly. An interacting multiple-model tracking algorithm based on multi-mode sensor scheduling is proposed for effectively tracking a smart target. Firstly, the motion of the smart target is modeled by the hybrid multiple models to describe its intelligent motion characteristics related to sensor modes. Secondly, based on the target-tracking error in one-step ahead, target motion-trend errors in two-step ahead, and the radiation cost of sensor mode, a multi-mode sensor scheduling strategy is proposed, and for which the object functions are built. Finally, the multi-mode sensor scheduling strategy and the interactive multiple model tracking algorithm are combined to form a cyclic closed-loop from sensor scheduling to interacting tracking and back to sensor scheduling. This closedloop operation effectively increases the adaptability in tracking a smart target. When compared with a fixed-mode sensor scheduling strategy or a multi-mode sensor scheduling with myopic policy, we find in simulations this sensor scheduling strategy is with higher accuracies in continuously tracking smart targets.