An efficient, variational approximation of the best fitting multi-Bernoulli filter
arXiv:1403.5022 · doi:10.1109/TSP.2014.2370946
Abstract
The joint probabilistic data association (JPDA) filter is a popular tracking methodology for problems involving well-spaced targets, but it is rarely applied in problems with closely-spaced targets due to its complexity in these cases, and due to the well-known phenomenon of coalescence. This paper addresses these difficulties using random finite sets (RFSs) and variational inference, deriving a highly tractable, approximate method for obtaining the multi-Bernoulli distribution that minimises the set Kullback-Leibler (KL) divergence from the true posterior, working within the RFS framework to incorporate uncertainty in target existence. The derivation is interpreted as an application of expectation-maximisation (EM), where the missing data is the correspondence of Bernoulli components (i.e., tracks) under each data association hypothesis. The missing data is shown to play an identical role to the selection of an ordered distribution in the same ordered family in the set JPDA algorithm. Subsequently, a special case of the proposed method is utilised to provide an efficient approximation of the minimum mean optimal sub-pattern assignment estimator. The performance of the proposed methods is demonstrated in challenging scenarios in which up to twenty targets come into close proximity.
Accepted, IEEE Transactions on Signal Processing, http://dx.doi.org/10.1109/TSP.2014.2370946
Cited by in corpus (25)
- Generalized optimal sub-pattern assignment metric
- Poisson multi-Bernoulli mixture filter: direct derivation and implementation
- Extended Object Tracking: Introduction, Overview and Applications
- Poisson multi-Bernoulli conjugate prior for multiple extended object filtering
- A metric on the space of finite sets of trajectories for evaluation of multi-target tracking algorithms
- Scalable Detection and Tracking of Geometric Extended Objects
- Trajectory PHD and CPHD filters
- A multi-sensor multi-Bernoulli filter
- Trajectory Poisson multi-Bernoulli filters
- Computationally Efficient Distributed Multi-sensor Fusion with Multi-Bernoulli Filter
- Distributed Multi-object Tracking under Limited Field of View Sensors
- A Poisson multi-Bernoulli mixture filter for coexisting point and extended targets
- Arithmetic Average Density Fusion -- Part II: Unified Derivation for Unlabeled and Labeled RFS Fusion
- A second-order PHD filter with mean and variance in target number
- A track-before-detect labelled multi-Bernoulli particle filter with label switching
- Poisson multi-Bernoulli mixture filter with general target-generated measurements and arbitrary clutter
- Data-driven clustering and Bernoulli merging for the Poisson multi-Bernoulli mixture filter
- Poisson Multi-Bernoulli Mixtures for Sets of Trajectories
- Multiple Object Trajectory Estimation Using Backward Simulation
- Arithmetic Average Density Fusion -- Part IV: Distributed Heterogeneous Fusion of RFS and LRFS Filters via Variational Approximation
- Spooky effect in optimal OSPA estimation and how GOSPA solves it
- Poisson Multi-Bernoulli Approximations for Multiple Extended Object Filtering
- Distributed Poisson multi-Bernoulli filtering via generalised covariance intersection
- Gaussian multi-target filtering with target dynamics driven by a stochastic differential equation
- Backward Simulation for Sets of Trajectories