A novel formulation of the Bayes recursion for single-cluster filtering

Brekke, Edmund; Kalyan, Bharath

doi:10.1109/aero.2014.6836493

Cited by 13 publications

(12 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [4], a generalized version of the PHD filter was used to solve the combined problem of SLAM and target tracking. A more general, but also more abstract, solution to the same problem was proposed in [5]. The simpler problem of detecting and tracking a single target using a moving platform was treated in [6].…”

Section: Previous Workmentioning

confidence: 99%

Suboptimal Kalman filters for target tracking with navigation uncertainty in one dimension

Brekke¹,

Wilthil²

2017

2017 IEEE Aerospace Conference

View full text Add to dashboard Cite

Abstract-The vast majority of literature on target tracking assumes that the position and orientation of the tracking sensor is stationary and/or known. However, for many applications the sensor is mounted on a moving vehicle, whose motion only can be estimated with a non-negligible uncertainty. In this paper, we suggest seven possible architectures for Kalman filtering in the simplest such scenario that we can construct: Assuming that both the ownship and a single target moves along a straight line according to linear kinematics. Some of the tracking filters are parameterized in the stationary world frame, while others are parameterized the body frame of the ownship. Also, some of the tracking filters take correlations between target state and ownship state into account, while others neglect such correlations. Simulations demonstrate that the suboptimal architectures may or may not reach similar performance as the optimal filter, depending on the process noise of the target and the performance measure chosen. Simulations of multitarget scenarios demonstrate that compensation of navigation uncertainty generally can reduce track-loss rates and OSPA distance.

show abstract

Section: Previous Workmentioning

confidence: 99%

Suboptimal Kalman filters for target tracking with navigation uncertainty in one dimension

Brekke¹,

Wilthil²

2017

2017 IEEE Aerospace Conference

View full text Add to dashboard Cite

show abstract

“…[6] and [12] in terms of Whypotheses. Under the given assumptions, the transition can also be expressed in terms of R-hypotheses.…”

Section: B the Predictionmentioning

confidence: 99%

“…For notational convenience, and in order to explore similarities with Reid's MHT, we use a matrix-based notation for this representation, originally proposed in [12]. The matrices are of dimension k×n k , where k is the last time step and n k is the number of measurements so far processed.…”

Section: A Williams' Hypothesesmentioning

confidence: 99%

“…Using the matrix-based formulation of W-hypotheses proposed in [12] (see Section III), these possibilities can be accounted for by partitioning the track indices i ∈ {1, . .…”

Section: The Posterior In Terms Of W-hypothesesmentioning

confidence: 99%

“…However, the relationship between FISST and Reid's formula has not yet been fully elucidated. This paper is inspired by the authors' previous work on MHT-techniques for simultaneous localization and mapping (SLAM) [11,12]. It borrows much notation from [12], which again was heavily influenced by preprints of [6] and the PhD thesis [13].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

The multiple hypothesis tracker derived from finite set statistics

Brekke¹

2017

2017 20th International Conference on Information Fusion (Fusion)

Self Cite

View full text Add to dashboard Cite

Abstract-The multiple hypothesis tracker (MHT) has historically been considered a gold standard for multi-target tracking. In this paper we show that the key formula for hypothesis probabilities in Reid's MHT can be derived from the modern theory of finite set statistics (FISST) insofar as appropriate assumptions (Poisson models for clutter and undetected targets, no target-death, linear-Gaussian Markov target kinematics) are adhered to.

show abstract