摘要
在多传感器数据融合系统中,航迹关联与系统误差估计之间存在着紧密的耦合关系。传感器系统误差容易诱发航迹关联出错;而系统误差的可靠估计又依赖于正确的关联结果。传统算法多忽视模块间耦合关系,对航迹关联与系统误差估计进行独立研究,在实际应用中性能退化严重。该文提出了一种基于稳健交替迭代的联合航迹关联与系统误差估计方法。该方法将错误关联视为系统误差估计的野值,在航迹关联与系统误差估计交替迭代过程中,使用最小平方中值(LMedS)估计器完成系统误差的稳健估计。仿真结果表明该方法在估计性能上具有明显优势。
Track association and sensor bias estimates in multi-sensor data fusion systems are mutually dependent on each other. Sensor biases can easily bring about misassociations, while the sensor bias estimates depend on correctly associated data. These two processes have been addressed separately but neglect of the coupling in real applications seriously degrades traditional algorithms. This paper presents a robust method for joint track association and sensor bias estimation. Misassociations were treated as special outliers in the sensor estimates with the least median of squares （LMedS） method used to obtain reliable bias estimates in alternating iterations of the track association and sensor bias estimation. Simulations show the advantages of this method.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第7期946-950,共5页
Journal of Tsinghua University(Science and Technology)
