Tracking Multiple Objects with Multimodal Dependent Measurements: Bayesian Nonparametric Modeling

We address the problem of multi-sensor multi-object tracking with unknown state parameter information. We develop algorithms capable of dealing with unknown time-dependent object and measurement cardinality and object identities. Additionally, given the dependent observations from sensors, our model takes advantage of the additional information to improve tracking performance. We robustly estimate the evolving objects and measurement cardinality with use of nonparametric modeling. In particular, we employ a dependent Dirichlet process to provide a prior on the time-varying object state distributions, and a hierarchical Dirichlet process (HDP) mixture to model measurement dependency. We demonstrate through simulations that providing multimodal dependent measurements the proposed method can accurately estimate the trajectory of objects and robustly determine the time-dependent cardinality. We also compare this method to the dependent Dirichlet process evolutionary Markov modeling (DDP-EMM) and demonstrate its improved tracking performance.

[1]  Steven Kay,et al.  Agile multi-modal tracking with dependent measurements , 2010, 2010 Conference Record of the Forty Fourth Asilomar Conference on Signals, Systems and Computers.

[2]  Antonia Papandreou-Suppappola,et al.  Random Infinite Tree and Dependent Poisson Diffusion Process for Nonparametric Bayesian Modeling in Multiple Object Tracking , 2019, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[3]  Eric P. Xing,et al.  The Dependent Dirichlet Process Mixture of Objects for Detection-free Tracking and Object Modeling , 2014, AISTATS.

[4]  Yee Whye Teh,et al.  Sharing Clusters among Related Groups: Hierarchical Dirichlet Processes , 2004, NIPS.

[5]  Muralidhar Rangaswamy,et al.  MULTI-TARGET TRACKING USING MULTI-MODAL SENSING WITH WAVEFORM CONFIGURATION , 2010 .

[6]  Antonia Papandreou-Suppappola,et al.  Use of Hierarchical Dirichlet Processes to Integrate Dependent Observations From Multiple Disparate Sensors for Tracking , 2019, 2019 22th International Conference on Information Fusion (FUSION).

[7]  Jonathan P. How,et al.  Dynamic Clustering via Asymptotics of the Dependent Dirichlet Process Mixture , 2013, NIPS.

[8]  Arnaud Doucet,et al.  Generalized Polya Urn for Time-varying Dirichlet Process Mixtures , 2007, UAI.

[9]  Bahman Moraffah,et al.  Inference for multiple object tracking: A Bayesian nonparametric approach , 2019, ArXiv.

[10]  Antonia Papandreou-Suppappola,et al.  Dependent Dirichlet Process Modeling and Identity Learning for Multiple Object Tracking , 2018, 2018 52nd Asilomar Conference on Signals, Systems, and Computers.

[11]  Michael I. Jordan,et al.  Bayesian Nonparametric Inference of Switching Dynamic Linear Models , 2010, IEEE Transactions on Signal Processing.

[12]  Antonia Papandreou-Suppappola,et al.  Nonparametric Bayesian Methods and the Dependent Pitman-Yor Process for Modeling Evolution in Multiple Object Tracking , 2019, 2019 22th International Conference on Information Fusion (FUSION).