| 摘要: |
| 为提高分布式雷达系统的目标检测与跟踪能力，研究了基于粒子滤波的检测前跟踪算法。针对传统粒子滤波中粗化方法盲目性的问题，提出了一种适用于分布式雷达目标检测与跟踪的多簇聚类粒子滤波算法。该算法在粗化的基础上，首先采用改进的K Means方法对粒子聚类以形成多个粒子簇，引导各簇内粒子沿着该簇中心向该簇最大联合似然粒子方向偏移，使粒子向高联合似然区域集中。该算法能够在缓解粒子滤波样本贫化问题的同时减少传统粗化的盲目性，提高了系统从接收数据中提取目标信息的能力。对分布式雷达目标检测与跟踪的仿真结果表明，多簇聚类粒子滤波算法比传统的粗化策略粒子滤波算法具有更好的检测能力和更高的跟踪精度。 |
| 关键词: 分布式雷达 检测前跟踪 粒子滤波 似然比 |
| DOI：DOI:10.3969/j.issn.1672-2337.2019.06.004 |
| 分类号:TN957.51 |
| 基金项目:国家自然科学基金（No.61701294） |
|
| Track Before Detect Algorithm for Distributed Radar Based on Particle Filter |
|
ZHANG Zhihong，ZHANG Kexin，PENG Zhangyou，ZHANG Zhonghao
|
|
Key Lab of Specialty Fiber Optics and Optical Access Networks，Shanghai University，Shanghai 200444，China
|
| Abstract: |
| In this paper，a track before detect algorithm based on particle filtering (PF) is proposed to improve the detection and tracking capability of distributed radar system.In specific，a new strategy based on clustering is introduced in our proposed algorithm to deal with the low efficiency of traditional roughening.In this strategy，all particles are first divided into a number of clusters via an improved K Means method. After that，for each cluster，the particles are moved toward the high joint likelihood region along a certain direction from the cluster center to the particle with maximum joint likelihood. In this way，one can cope with the problem of sample impoverishment more efficiently than the traditional roughening strategy. The simulation results are presented to demonstrate that our proposed PF algorithm outperforms the standard PF algorithm with traditional roughening in target detection and tracking for distributed radar system. |
| Key words: distributed radar track before detect particle filter likelihood ratio |
