Abstract: A federated fusion algorithm of joint multi-Gaussian mixture multi-Bernoulli (JMGM-MB) filters is proposed to achieve optimal fusion tracking of multiple uncertain maneuvering targets in a hierarchical structure. The JMGM-MB filter achieves a higher level of accuracy than the multi-model Gaussian mixture MB (MM-GM-MB) filter by propagating the state density of each potential target in the interactive multi-model (IMM) filtering manner. Within the hierarchical structure, each sensor node performs a local JMGM-MB filter to capture survival, newborn, and vanishing targets. A notable characteristic of our algorithm is a master filter running on the fusion node, which can help identify the origins of state estimates and supplement missed detections. The outputs of all filters are associated into multiple groups of single-target estimates. We rigorously derive the optimal fusion of IMM filters and apply it to merge associated single-target estimates. This optimality is guaranteed by the covariance upper-bounding technique, which can truly eliminate correlations among filters. Simulation results demonstrate that the proposed algorithm outperforms the existing centralized and distributed fusion algorithms in linear and heterogeneous scenarios, and the relative weights of the master and local filters can be adjusted flexibly.

基于多伯努利滤波器的多机动目标最优联邦融合

[1]Balenzuela MP, Wills AG, Renton C, et al., 2022. Parameter estimation for jump Markov linear systems. Automatica, 135:109949.

[2]Carlson NA, 1990. Federated square root filter for decentralized parallel processors. IEEE Trans Aerosp Electron Syst, 26(3):517-525.

[3]Chang CB, Athans M, 1978. State estimation for discrete systems with switching parameters. IEEE Trans Aerosp Electron Syst, AES-14(3):418-425.

[4]Da K, Li TC, Zhu YF, et al., 2020a. Gaussian mixture particle jump-Markov-CPHD fusion for multitarget tracking using sensors with limited views. IEEE Trans Signal Inform Process Netw, 6:605-616.

[5]Da K, Li TC, Zhu YF, et al., 2020b. Kullback‍–Leibler averaging for multitarget density fusion. Proc 16^th Int Symp on Distributed Computing and Artificial Intelligence, p.253-261.

[6]Dong XD, Zhang XF, Zhao J, et al., 2021. Multi-maneuvering sources DOA tracking with improved interactive multi-model multi-Bernoulli filter for acoustic vector sensor (AVS) array. IEEE Trans Veh Technol, 70(8):7825-7838.

[7]Dunne D, Kirubarajan T, 2013. Multiple model multi-Bernoulli filters for manoeuvering targets. IEEE Trans Aerosp Electron Syst, 49(4):2679-2692.

[8]Gao L, Battistelli G, Chisci L, 2020. Multiobject fusion with minimum information loss. IEEE Signal Process Lett, 27:201-205.

[9]Georgescu R, Willett P, 2012. The multiple model CPHD tracker. IEEE Trans Signal Process, 60(4):1741-1751.

[10]Gunay M, Orguner U, Demirekler M, 2016. Chernoff fusion of Gaussian mixtures based on sigma-point approximation. IEEE Trans Aerosp Electron Syst, 52(6):2732-2746.

[11]Hu GG, Xu LY, Gao BB, et al., 2023. Robust unscented Kalman filter-based decentralized multisensor information fusion for INS/GNSS/CNS integration in hypersonic vehicle navigation. IEEE Trans Instrum Meas, 72:8504011.

[12]Hu XL, Zhang Q, Song BJ, et al., 2022. σ-threshold Bayes filter in unknown birth background with multi-Bernoulli finite sets. Proc IEEE Int Conf on Signal Processing, Communications and Computing, p.1-4.

[13]Julier SJ, Uhlmann JK, 1997. A non-divergent estimation algorithm in the presence of unknown correlations. Proc American Control Conf, p.2369-2373.

[14]Li GY, Battistelli G, Chisci L, et al., 2024. Distributed joint detection, tracking, and classification via labeled multi-Bernoulli filtering. IEEE Trans Cybern, 54(3):1429-1441.

[15]Li TC, 2024. Arithmetic average density fusion—part II: unified derivation for unlabeled and labeled RFS fusion. IEEE Trans Aerosp Electron Syst, 60(3):3255-3268.

[16]Li TC, Corchado JM, Chen HM, 2018. Distributed flooding-then-clustering: a lazy networking approach for distributed multiple target tracking. Proc 21^st Int Conf on Information Fusion, p.2415-2422.

[17]Li TC, Corchado JM, Sun SD, 2019a. Partial consensus and conservative fusion of Gaussian mixtures for distributed PHD fusion. IEEE Trans Aerosp Electron Syst, 55(5):‍2150-2163.

[18]Li TC, Fan HQ, García J, et al., 2019b. Second-order statistics analysis and comparison between arithmetic and geometric average fusion: application to multi-sensor target tracking. Inform Fus, 51:233-243.

[19]Li TC, Wang XX, Liang Y, et al., 2020. On arithmetic average fusion and its application for distributed multi-Bernoulli multitarget tracking. IEEE Trans Signal Process, 68:‍2883-2896.

[20]Li TC, Song Y, Song EB, et al., 2024a. Arithmetic average density fusion—part I: some statistic and information-theoretic results. Inform Fus, 104:102199.

[21]Li TC, Yan RB, Da K, et al., 2024b. Arithmetic average density fusion—part III: heterogeneous unlabeled and labeled RFS filter fusion. IEEE Trans Aerosp Electron Syst, 60(1):‍‍‍1023-1034.

[22]Li WL, Jia YM, 2011. Gaussian mixture PHD filter for jump Markov models based on best-fitting Gaussian approximation. Signal Process, 91(4):1036-1042.

[23]Mahler R, 2014. Advances in Statistical Multisource-Multitarget Information Fusion. Artech House, Norwood, USA.

[24]Ouyang C, Ji HB, Guo ZQ, 2012. Extensions of the SMC-PHD filters for jump Markov systems. Signal Process, 92(6):1422-1430.

[25]Peng C, Chai L, Yi W, et al., 2021. Distributed multi-sensor multi-view fusion of PHD filter for maneuvering targets. Proc CIE Int Conf on Radar, p.1182-1187.

[26]Punithakumar K, Kirubarajan T, Sinha A, 2008. Multiple-model probability hypothesis density filter for tracking maneuvering targets. IEEE Trans Aerosp Electron Syst, 44(1):87-98.

[27]Schuhmacher D, Vo BT, Vo BN, 2008. A consistent metric for performance evaluation of multi-object filters. IEEE Trans Signal Process, 56(8):3447-3457.

[28]Shen-Tu H, Lin RF, Shen WC, et al., 2024. An arithmetic geometric mixed average GM-PHD algorithm for decentralized sensor network with limited field of view. IEEE Sens J, 24(12):19995-20008.

[29]Sun SL, 2020. Distributed optimal linear fusion estimators. Inform Fus, 63:56-73.

[30]Sun YC, Kim D, Hwang I, 2022. Multiple-model Gaussian mixture probability hypothesis density filter based on jump Markov system with state-dependent probabilities. IET Radar Sonar Navig, 16(11):1881-1894.

[31]Üney M, Clark DE, Julier SJ, 2013. Distributed fusion of PHD filters via exponential mixture densities. IEEE J Sel Top Signal Process, 7(3):521-531.

[32]Üney M, Houssineau J, Delande E, et al., 2019. Fusion of finite-set distributions: pointwise consistency and global cardinality. IEEE Trans Aerosp Electron Syst, 55(6):2759-2773.

[33]Vo BT, Vo BN, Cantoni A, 2009. The cardinality balanced multi-target multi-Bernoulli filter and its implementations. IEEE Trans Signal Process, 57(2):409-423.

[34]Wang KW, Zhang Q, Hu XL, 2024. Multisensor multitarget tracking arithmetic average fusion method based on probabilistic time window. IEEE Sens J, 24(3):3583-3593.

[35]Wei JX, Luo F, Chen SC, et al., 2023. Robust fusion of GM-PHD filters based on geometric average. Signal Process, 206:108912.

[36]Wu SY, Dong XD, Zhao J, et al., 2019. A fast implementation of interactive-model generalized labeled multi-Bernoulli filter for interval measurements. Signal Process, 164:‍345-353.

[37]Wu WH, Cai YC, Jin HB, et al., 2021a. Derivation of the multi-model generalized labeled multi-Bernoulli filter: a solution to multi-target hybrid systems. Front Inform Technol Electron Eng, 22(1):79-87.

[38]Wu WH, Sun HM, Huang ZL, et al., 2021b. Multi-GMTI fusion for Doppler blind zone suppression using PHD fusion. Signal Process, 183:108024.

[39]Xie X, Sun HM, Wu WH, et al., 2019. Multi-UAV multi-target tracking in the presence of Doppler blind zone. Proc IEEE Int Conf on Unmanned Systems, p.438-442.

[40]Xie XX, Wang Y, Guo JQ, et al., 2023. The multiple model Poisson multi-Bernoulli mixture filter for extended target tracking. IEEE Sens J, 23(13):14304-14314.

[41]Xue Y, Feng XA, 2024. Joint multi-Gaussian mixture model and its application to multi-model multi-Bernoulli filter. Dig Signal Process, 153:104616.

[42]Yang H, Li TC, Yan JK, et al., 2024. Hierarchical average fusion with GM-PHD filters against FDI and DoS attacks. IEEE Signal Process Lett, 31:934-938.

[43]Yi W, Li SQ, Wang BL, et al., 2020. Computationally efficient distributed multi-sensor fusion with multi-Bernoulli filter. IEEE Trans Signal Process, 68:241-256.

[44]Zhao BF, 2024. Multisensor maneuvering target fusion tracking using interacting multiple model. Autom Contr Comput Sci, 58(3):303-312.

[45]Zhou YQ, Yan LP, Li H, et al., 2024. The multiple pairwise Markov chain model-based labeled multi-Bernoulli filter. J Franklin Inst, 361(10):106939.