| 摘要: |
| 针对组网雷达系统误差与上报目标不完全一致等复杂环境下的航迹关联鲁棒性问题，基于非刚性点集配准理论提出了一种航迹邻域结构信息与拉普拉斯混合模型(Laplace Mixture Model,LMM)相结合的关联方法。首先利用更鲁棒的拉普拉斯混合模型对被视为异常点的非共同探测目标的航迹进行建模，然后定义局部相似性测度计算航迹邻域结构的相似性来决定拉普拉斯分量的权重，并通过期望最大化(Expectation-Maximization,EM)算法求解拉普拉斯混合模型的闭合解。最后利用经典分配法对E步获取的后验概率矩阵进行航迹关联判决。仿真结果表明，该方法在面对各种复杂环境如不同系统误差、检测概率、目标分布密度等情况时均有较高的关联正确率和较好的鲁棒性。 |
| 关键词: 组网雷达系统 航迹关联 拉普拉斯混合模型 局部相似性测度 EM算法 |
| DOI：10.20079/j.issn.1001-893x.230927002 |
|
| 基金项目:国家自然科学基金资助项目(62371149,62263007);广西科技重大专项(AA20302001);广西无线宽带通信与信号处理重点实验室基金(GXKL06190117);认知无线电与信息处理教育部重点实验室基金(CRKL180106,CRKL220107,CRKL210101);桂林电子科技大学校级研究生创新项目(2023YCXS108) |
|
| An Anti-bias Track Association Algorithm Based on Laplace Mixture Model |
| WEI Chunlinga,WU Sunyongb,LIU Jinxina,YU Runhuaa |
| (1a.School of Information and Communication;1b.School of Mathematics and Computing Science,Guilin University of Electronic Technology,Guilin 541004,China;2.Ministry of Education Key Laboratory of Cognitive Radio and Information Processing,Guilin 541004,China) |
| Abstract: |
| In order to solve the robustness issue of track association in complex environments with system errors and incomplete consistency with reported targets in networked radar system,according to the non-rigid point set registration theory,an association method which combines the neighborhood structural information of the track and the Laplace Mixture Model(LMM) is proposed.The trajectories of non-cooperative detection targets,deemed as outliers,are initially modeled by a more robust LMM.Then,a local similarity measurement is defined to calculate the similarity of track neighborhood structures to determine the weights of Laplace components,and the closed-form solution of the LMM is obtained through the Expectation-Maximization(EM) algorithm.Finally,a classical assignment method is employed to make track association decisions based on the posterior probability matrix obtained from the expectation step.Simulation results demonstrate that the proposed algorithm achieves high association accuracy and robustness when confronted with various complex environments,such as different system errors,detection probabilities,and target distribution densities. |
| Key words: networked radar system track association Laplace mixture model local similarity measurement EM algorithm |
