摘要
多目标跟踪是多传感器系统信息融合中的核心技术之一。采用新型的AFCM模糊算法实现对多目标交叉状态下航迹数据关联。该算法定义了一种新的度量空间中的距离,通过新的距离定义有效抑制含有噪声点的样本及目标航迹交叉在迭代中对数据关联聚类中心点的大幅偏差。同时应用改进带加权的航迹融合算法对红外和毫米波雷达传感器测量的航迹数据进行融合。仿真试验证明,新的算法在综合多传感器探测优势的基础上,对航迹的融合结果优于SF算法。新的数据关联算法和改进的加权航迹融合算法为多源信息融合提供了一种可靠有效的多目标跟踪技术。
Multi-target tracking is one of the core technologies in the multi-sensor information fusion. A novel AFCM algorithm is proposed aiming at realizing the multi-target tracking under the cross tracks situation. This new algorithm defines a new distance in new metric space, meanwhile using the improved weighted track fusion algorithm to realize IR and MMW radar sensors＇ track fusion. The new metric has been properly verified and the clustering output improves the data association. Experiments have proved that the novel system has an outstanding performance over normal system from considering the advantages of the multi-sensor and the process complexity. The application of the AFCM optimizes the process of the track fusion and provides a reliable method to realize the multi-target tracking.
出处
《传感技术学报》
CAS
CSCD
北大核心
2009年第3期358-361,共4页
Chinese Journal of Sensors and Actuators
基金
航天创新基金资助
航天支撑基金
国家自然科学基金项目支持(90306008)
