自适应转移概率的IMM-FPF算法
|
摘要
结合反馈粒子滤波(Feedback Particle Filter,FPF)和交互式多模型(Interacting multiple model,IMM)滤波算法,提出了一种自适应转移概率的机动目标跟踪IMM-FPF算法。该算法利用当前量测信息,实时更新Markov转移概率矩阵,有效的克服了传统的IMM算法中转移概率先验已知的限定,改善了模型跟踪精度和稳定性;同时将FPF算法的实现应用于线性离散系统,然后对各个模型随机采样相同数目的粒子,经过输入交互、反馈粒子滤波后,再进行估计融合。由于在滤波过程中不需要对粒子进行重采样和分类,因此在保证算法跟踪精度的同时,减小了计算量,提高了算法的实时性。仿真结果表明,该算法的实时性及其跟踪性能均优于交互式多模型粒子滤波(IMM-PF)算法。
The IMM-FPF algorithm of the adaptive transition probability matrix is proposed for maneuvering target tracking in this paper,based on feedback particle filter(FPF)and interacting multiple models(IMM).By using the current measurements,Markov transition probability matrix can be updated.So the proposed algorithm effectively removes the limit that a determination priori of the transition probability is known in the traditional IMM,which improves the tracking accuracy and stability.At the same time,the implementation of feedback particle filter is introduced to the linear discrete system.The initial particles of same number are sampled randomly for every model.After interaction and feedback particle filtering,the particles are used for estimation fusion.Because it is not necessary to resample the particles in the update,the amount of calculation is reduced effectively at the same time to ensure the accuracy of the algorithm.Finally,the simulation results show that the tracking and real-time performance of the proposed method can been improved,compared with the standard IMM-PF.
The IMM-FPF algorithm of the adaptive transition probability matrix is proposed for maneuvering target tracking in this paper,based on feedback particle filter(FPF)and interacting multiple models(IMM).By using the current measurements,Markov transition probability matrix can be updated.So the proposed algorithm effectively removes the limit that a determination priori of the transition probability is known in the traditional IMM,which improves the tracking accuracy and stability.At the same time,the implementation of feedback particle filter is introduced to the linear discrete system.The initial particles of same number are sampled randomly for every model.After interaction and feedback particle filtering,the particles are used for estimation fusion.Because it is not necessary to resample the particles in the update,the amount of calculation is reduced effectively at the same time to ensure the accuracy of the algorithm.Finally,the simulation results show that the tracking and real-time performance of the proposed method can been improved,compared with the standard IMM-PF.
引文
[1] Mazor E,Averbuch A,Bar-Shalom Y,et al.Interacting multiple model methods in target tracking:a survey[J].IEEE Trans.on Aerospace and Electronic Systems,1998,34(1):103-123.
[2] Brehard T,Cadre J P.Distributed target tracking for nonlinear systems:Application to bearings-only tracking[C].The 7th Int.Conf.on Information Fusion,Philadelphia,2005:77-84.
[3] Gordon N J,Salmond D J and Smith A F.M.Novel approach to nonlinear/non-Gaussian Bayesian state estimation[J].IEE Proceedings-F,Radar and Signal Process,1993,140(2):107-113.
[4] Boers Y,Driessen J N.Interacting multiple model particle filter[J].IET Proceedings Radar Sonar and Navigation,2003,150(5):334-349.
[5]邓小龙,谢剑英,杨煜普.基于交互式多模型的粒子滤波算法[J].系统仿真学报,2005,17(10):2360-2363.
[6] Bugallo M F,Xu S S,Petar M D.Performance comparison of EKF and particle filtering methods for maneuvering targets[J].Digital Signal Process,2006,16(10):67-78.
[7]彭志专,冯金富,钟咏兵,等.基于IMM-PF的分布式估计融合算法[J].控制与决策,2008,23(7):837-840.
[8]曹洁,文如泉.IMM-UPF算法在机动目标跟踪中的研究[J].计算机工程与应用,2010,46(28):240-243.
[9]张俊根,姬红兵.IMM迭代扩展卡尔曼粒子滤波跟踪算法[J].电子与信息学报,2010,32(5):111-1120.
[10] Yang T.,Blom H.A.P.,Mehta,P.G.Interacting multiple model-feedback particle filter for stochastic hybrid systems[C].IEEE 52nd Annual Conference on Decision and Control,Florence,Italy,Dec.,2013,pp:7065-7070.
[11] Doucet A,Ristic,B. Recursive state estimation for multiple switching models with unknown transition probabilities[J].IEEE Trans.on Aerospace and Electronic Systems,2002,38(3):1098-1104.
[12] Yang C Y,Chen B S,Liao F K.Mobile location estimation using fuzzy-based IMM and data fusion[J].IEEE Trans.on Mobile Computing,2010,9(10):1424-1436.
[13]黄鹤,王小旭,赵春晖,等.基于后验信息修正的自适应交互多模型跟踪算法[J].西北工业大学学报,2011,29(6):829-833.
[14]封普文,黄长强,曹林平,等.马尔可夫矩阵修正IMM跟踪算法[J].系统工程与电子技术,2013,35(11):2269-2274.
[15]郭志,董春云,蔡远利,等.时变转移概率IMM-SRCKF机动目标跟踪算法[J].系统工程与电子技术,2015,37(1):24-30.
[16] Yang T,Mehta P G,Meyn S P.Feedback particle filter[J].IEEE Trans.on Automatic Control,2013,58(10):2465-2480.
[17]秦岭、刘晨曦.反馈粒子滤波在GPS/INS组合导航系统中的应用[J].武汉轻工大学学报,2015,34(4):56-59.
[18] Yang T,Laugesen R S,Mehta P G.,et.al.Multivariable feedback particle filter[C].Proc.51st IEEE Conf.Decision Control,Maui,HI,USA,Dec,2012,pp:4063-4070.
[19] Tilton A K,Hsiao-Wecksler E T,Mehta P G.Filtering with rhythms:Application of the feedback particle filter to estimation of human gait cycle[C].Proc.2012American Control Conf.,Montréal,QC,Canada,Jun,2012,pp:3433-3438.
[20] Tilton A K,Mehta P G,Meyn S P.Multi-dimensional feedback particle filter for coupled oscillators[C].Proc.2013 American Control Conf.,Washington,DC,Jun,2013,pp:2415-2421.
[2] Brehard T,Cadre J P.Distributed target tracking for nonlinear systems:Application to bearings-only tracking[C].The 7th Int.Conf.on Information Fusion,Philadelphia,2005:77-84.
[3] Gordon N J,Salmond D J and Smith A F.M.Novel approach to nonlinear/non-Gaussian Bayesian state estimation[J].IEE Proceedings-F,Radar and Signal Process,1993,140(2):107-113.
[4] Boers Y,Driessen J N.Interacting multiple model particle filter[J].IET Proceedings Radar Sonar and Navigation,2003,150(5):334-349.
[5]邓小龙,谢剑英,杨煜普.基于交互式多模型的粒子滤波算法[J].系统仿真学报,2005,17(10):2360-2363.
[6] Bugallo M F,Xu S S,Petar M D.Performance comparison of EKF and particle filtering methods for maneuvering targets[J].Digital Signal Process,2006,16(10):67-78.
[7]彭志专,冯金富,钟咏兵,等.基于IMM-PF的分布式估计融合算法[J].控制与决策,2008,23(7):837-840.
[8]曹洁,文如泉.IMM-UPF算法在机动目标跟踪中的研究[J].计算机工程与应用,2010,46(28):240-243.
[9]张俊根,姬红兵.IMM迭代扩展卡尔曼粒子滤波跟踪算法[J].电子与信息学报,2010,32(5):111-1120.
[10] Yang T.,Blom H.A.P.,Mehta,P.G.Interacting multiple model-feedback particle filter for stochastic hybrid systems[C].IEEE 52nd Annual Conference on Decision and Control,Florence,Italy,Dec.,2013,pp:7065-7070.
[11] Doucet A,Ristic,B. Recursive state estimation for multiple switching models with unknown transition probabilities[J].IEEE Trans.on Aerospace and Electronic Systems,2002,38(3):1098-1104.
[12] Yang C Y,Chen B S,Liao F K.Mobile location estimation using fuzzy-based IMM and data fusion[J].IEEE Trans.on Mobile Computing,2010,9(10):1424-1436.
[13]黄鹤,王小旭,赵春晖,等.基于后验信息修正的自适应交互多模型跟踪算法[J].西北工业大学学报,2011,29(6):829-833.
[14]封普文,黄长强,曹林平,等.马尔可夫矩阵修正IMM跟踪算法[J].系统工程与电子技术,2013,35(11):2269-2274.
[15]郭志,董春云,蔡远利,等.时变转移概率IMM-SRCKF机动目标跟踪算法[J].系统工程与电子技术,2015,37(1):24-30.
[16] Yang T,Mehta P G,Meyn S P.Feedback particle filter[J].IEEE Trans.on Automatic Control,2013,58(10):2465-2480.
[17]秦岭、刘晨曦.反馈粒子滤波在GPS/INS组合导航系统中的应用[J].武汉轻工大学学报,2015,34(4):56-59.
[18] Yang T,Laugesen R S,Mehta P G.,et.al.Multivariable feedback particle filter[C].Proc.51st IEEE Conf.Decision Control,Maui,HI,USA,Dec,2012,pp:4063-4070.
[19] Tilton A K,Hsiao-Wecksler E T,Mehta P G.Filtering with rhythms:Application of the feedback particle filter to estimation of human gait cycle[C].Proc.2012American Control Conf.,Montréal,QC,Canada,Jun,2012,pp:3433-3438.
[20] Tilton A K,Mehta P G,Meyn S P.Multi-dimensional feedback particle filter for coupled oscillators[C].Proc.2013 American Control Conf.,Washington,DC,Jun,2013,pp:2415-2421.