摘要
为了解决马尔科夫跳变系统广义标签多伯努利滤波器在多机动目标跟踪场景需要计算大量模式假设分支,并且需要频繁对假设分支进行剪枝,导致算法存在计算量过高并且影响跟踪精度的问题,提出一种基于交互多模型的分组δ-广义标签多伯努利滤波器。滤波器采用航迹分组策略,不同组的航迹独立进行关联映射与分支权重计算,降低了关联的计算复杂度,可以实现不同航迹组之间并行滤波。另外,为了处理机动目标场景引入交互多模型,给出基于交互多模型的分组滤波递推方程。仿真结果表明,所提出的滤波器跟踪精度更高,计算速度更快,可以用于跟踪多个机动目标的场景。
When tracking multi-maneuvering targets with jump Markov system generalized labeled multi-Bernoulli filter,there are too many motion mode hypotheses that need to be calculated and pruned frequently,which may increase the computational complexity and negatively affect the tracking accuracy.In order to solve the problem,an interacting multiple model(IMM)based groupingδ-generalized labeled multi-Bernoulli filter is proposed.With this filter,all tracks fall into different groups,association mapping and hypothesis weight calculation run in each individual group,which reduces the computational complexity and enables parallelization.Moreover,IMM algorithm is incorporated into this filter to deal with target maneuver.The time prediction and data update equations are given in detail.Simulation results show that the proposed filter can track multiple maneuvering targets with higher accuracy and lower computational cost.
作者
辛怀声
曹晨
XIN Huaisheng;CAO Chen(China Academy of Electronics and Information Technology, Beijing 100041)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2022年第4期1128-1138,共11页
Systems Engineering and Electronics
关键词
随机有限集
δ-广义标签多伯努利滤波器
多目标跟踪
交互多模型
random finite set(RFS)
δ-generalized labeled multi-Bernoulli(δ-GLMB)filter
multi target tracking(MTT)
interacting multiple mode(IMM)
