系统工程与电子技术 ›› 2025, Vol. 47 ›› Issue (7): 2086-2097.doi: 10.12305/j.issn.1001-506X.2025.07.02

• 电子技术 • 上一篇    

战术机动先验辅助的交互多模型目标跟踪方法

孙一心, 方愚渊, 张磊   

  1. 中山大学电子与通信工程学院, 广东 深圳 518107
  • 收稿日期:2024-06-08 出版日期:2025-07-16 发布日期:2025-07-22
  • 通讯作者: 张磊
  • 作者简介:孙一心(2001—), 女, 硕士研究生, 主要研究方向为雷达认知探测
    方愚渊(1994—), 男, 博士研究生, 主要研究方向为雷达认知抗干扰、协同抗干扰
    张磊(1984—), 男, 教授, 博士, 主要研究方向为新体制雷达、雷达成像与识别感知
  • 基金资助:
    国家自然科学基金(62101603);国家自然科学基金(62201623)

Tactical maneuver priori assisted interactive multiple model target tracking method

Yixin SUN, Yuyuan FANG, Lei ZHANG   

  1. School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen 518107, China
  • Received:2024-06-08 Online:2025-07-16 Published:2025-07-22
  • Contact: Lei ZHANG

摘要:

针对交互式多模型滤波算法应对强机动目标跟踪过程存在响应滞后、精度偏低的问题, 提出由空中目标战术机动先验辅助的层次性交互多模型跟踪方法。所提方法利用隐马尔可夫模型, 对目标机动轨迹分段进行战术类别的识别、战术状态的区分, 建模不同战术下的机动模型转移矩阵。由此实现战术机动似然概率与交互多模型机动似然概率的有效交互, 构建双层交互式多模型跟踪算法。实验考虑复杂战术意图下空中目标机动轨迹, 结果表明所提算法结合战术机动先验可有效解决状态估计精度与机动转移响应时效之间的矛盾, 明显改善机动目标跟踪精度。

关键词: 交互式多模型, 马尔可夫转移概率, 机动目标跟踪, 先验信息

Abstract:

To address the issue of response lag and low accuracy in tracking highly maneuvering targets with the interactive multiple model (IMM) filter algorithm, a hierarchical IMM tracking method assisted by tactical maneuver priors of aerial targets is proposed. The proposed method employs a hidden Markov model (HMM) to segment and classify the tactical categories of target maneuver trajectories and to distinguish between tactical states, and models the transfer matrix of maneuver models with different tactics. This enables the effective interaction of tactical likelihood and the IMM maneuvering models' likelihood, resulting in the development of a double-layered IMM (DIMM) tracking algorithm. The experiments consider the air target maneuvering trajectories with complex tactical intentions, and result shows that the proposed algorithm combined with the tactical maneuver priori effectively solves the contradiction between state estimation accuracy and maneuver transition response time, and significantly improves the maneuvering target tracking accuracy.

Key words: interactive multiple model (IMM), Markov transition probability, maneuvering target tracking, prior information

中图分类号: