基于双马尔可夫链的SMC-CBMeMBer滤波

大部分多目标跟踪滤波器都是假设目标及其量测符合隐式马尔可夫链(hidden Markov chain,HMC)模型,而HMC模型隐含的独立性假定在很多实际应用中是无效的,双马尔可夫链(pairwise Markov chain,PMC)模型相对于HMC模型更具有普适性。已有的基于PMC模型的势均衡多目标多伯努利(cardinality balanced multi-target multi-Bernoulli,CBMeMBer)滤波的高斯混合实现仅适用于线性高斯系统,针对基于PMC模型的非线性多目标跟踪系统,将每一条假设航迹的伯努利随机有限集用一组加权粒子来近似,提出了基于PMC模型的势均衡多目标多伯努利滤波的序贯蒙特卡罗(sequential Monte-Carlo,SMC)方法实现(SMC-PMC-CBMeMBer)滤波。仿真实验结果验证了SMC-PMC-CBMeMBer算法的有效性,在基于PMC模型的非线性多目标跟踪系统中,SMC-PMC-CBMeMBer算法性能优于基于HMC模型的SMC-CBMeMBer滤波器和基于PMC模型的SMC-PHD滤波器。

SMC-CBMeMBer filter based on pairwise Markov chains

Most multi target tracking filters assume that one target and its observation follow a hidden Markov chain (HMC) model, but the implicit independence assumption of the HMC model is invalid in many practical applications, and a pairwise Markov chain (PMC) model is more universally suitable than the traditional HMC model. The existing Gauss mixture implementation of cardinality balanced multi-target multi-Bernoulli (CBMeMBer) filter based on the PMC model is only applicable to the linear Gauss system.Each hypothetical path of the Bernoulli random finite set is approximated by a set of weighted particles, and then the sequential Monte-Carlo(SMC) implementation of the PMC-CBMeMBer filter is proposed for nonlinear systems. The experimental results show that SMC-PMC-CBMeMBer filter has better tracking performance than the SMC-HMC-CBMeMBer filter and the SMC-PMC-PHD filter.