基于高斯过程的航迹片段关联算法

Track Segment Association Algorithm Based on Gaussian Process

在恶劣战场环境下,目标航迹极易中断并形成大量零碎片段,对航迹管理和态势估计产生了严峻挑战。传统的航迹片段关联算法在假设的先验模型和实际运动模式不匹配时,性能大幅下降。针对这个问题,提出一种基于高斯过程的航迹片段关联算法。利用高斯过程对目标的状态转移函数非参数学习,通过对中断区间的量测进行训练将航迹新旧片段分别回溯和预测至关联时刻。采用假设检验和二维分配技术对航迹片段进行关联和配对。在整个中断区间,通过高斯过程将配对后的新旧航迹进行航迹缝合。仿真结果表明,在目标先验模型与实际运动模式不匹配时,所提算法能有效提高航迹正确关联率和航迹寿命。

The target track is easily broken and forms a large number of segments in the harsh battlefield environment,which poses a severe challenge to track management and situation estimation.The performance of traditional track segment association algorithm is greatly reduced when the assumed prior model does not match the actual motion pattern.To solve this problem,a track segment association algorithm based on Gaussian process is proposed.First,Gaussian process is used to make a nonparametric model learning of the state transition function of the target,and the new and old segments of the track are traced back and predicted to the association time respectively by training the measurement of the interrupted zone.Second,hypothesis testing and two-dimensional assignmen technologyt are used to associate and pair matching of the track segments.Finally,in the whole interrupted zone,the paired new and old track segments are stitched by Gaussian process.The simulation results show that the proposed algorithm can effectively improve the track correct association rate and track lifetime when the target prior model does not match the actual motion pattern.