Multi-Scan Multi-Sensor Multi-Object State Estimation

Moratuwage, Diluka; Vo, Ba-Ngu; Vo, Ba-Tuong; Shim, Changbeom

doi:10.1109/tsp.2022.3218366

Cited by 19 publications

(5 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the effectiveness of the proposed approach, extension to the multi-dimensional assignment problem [37] could alleviate the computational bottlenecks in multi-scan and/or multisensor truncation. The recent multi-sensor multi-scan GLMB smoother proposed in [12] extends the SGS solution to the multi-dimensional assignment problem. While this is the first solution to address multi-dimension assignment problems of such large scale, we envisage that its time complexity can be drastically reduced using our proposed TGS approach.…”

Section: Discussionmentioning

confidence: 99%

“…Noting that P (the number of hypothesized objects) is strongly correlated with M (the number of measurements), this complexity translates roughly to a cubic complexity, i.e., O(T P 3 ) or O(T M 3 ). Nonetheless, SGS has been extended to address multi-dimensional ranked assignment problems in multi-scan and multi-sensor GLMB filtering [8], [11], [12].…”

Section: B Gibbs Sampling For Glmb Truncationmentioning

confidence: 99%

“…Moreover, the GLMB family of densities is furnished with elegant mathematical properties, offering a versatile set of tools including, conjugacy [7], closure under truncation with analytic truncation error [8], analytic approximation of general labeled multi-object density with minimal Kullback-Leibler divergence [9], analytic Cauchy-Schwarz divergence and void probabilities [10]. These properties enabled MOT with multiple sensors and scans [8], [11], [12], non-standard models [9], [13], [14], [15], unknown system parameters [16], multi-object control solutions [10], [17], as well as distributed implementations [18], [19], [20]. The GLMB filter is also amenable to parallelization that reduces computation times.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Linear Complexity Gibbs Sampling for Generalized Labeled Multi-Bernoulli Filtering

Shim

et al. 2023

IEEE Trans. Signal Process.

Self Cite

View full text Add to dashboard Cite

Generalized Labeled Multi-Bernoulli (GLMB) densities arise in a host of multi-object system applications analogous to Gaussians in single-object filtering. However, computing the GLMB filtering density requires solving NP-hard problems. To alleviate this computational bottleneck, we develop a linear complexity Gibbs sampling framework for GLMB density computation. Specifically, we propose a tempered Gibbs sampler that exploits the structure of the GLMB filtering density to achieve an O(T (P + M )) complexity, where T is the number of iterations of the algorithm, P and M are the number hypothesized objects and measurements. This innovation enables the GLMB filter implementation to be reduced from an O(T P 2 M ) complexity to O(T (P + M + log T ) + P M). Moreover, the proposed framework provides the flexibility for trade-offs between tracking performance and computational load. Convergence of the proposed Gibbs sampler is established, and numerical studies are presented to validate the proposed GLMB filter implementation.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: B Gibbs Sampling For Glmb Truncationmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Linear Complexity Gibbs Sampling for Generalized Labeled Multi-Bernoulli Filtering

Shim

et al. 2023

IEEE Trans. Signal Process.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Regardless of some recent definitions of disvergence/distances for two distributions of different dimensions or domains [35], there still lacks a proper distance/divergence for the LRFS densities with different, discrete labels. As such, both (14) and (15) can not be directly applied for the labeled RFS density distributions in general. Yet, this has been often violated in the literature.…”

Section: Phd-aa Consistencymentioning

confidence: 99%

“…This needs an inter-node communication protocol like flooding [6] which is often communication costly and does not suit the large-scale peer-to-peer distributed networks, namely unscalable with the number of sensors. More importantly, it is computationally intractable for the RFS filters to make the optimal use of the measurements of multiple sensors due to the explosive possibility for track-measurement association [4], [7]- [15]. In this paper, we resort to the computationally efficient, scalable, robust density fusion approach [16] to the coordination of the probability hypothesis density (PHD) filter [17], [18], the unlabeled multiple Bernoulli (MB) filter [19] and the labeled MB (LMB) filter [20], [21].…”

Section: Introductionmentioning

confidence: 99%

Distributed Multisensor Multitarget Tracking Algorithm with Time-Offset Registration

Chen

Wang

et al. 2020

JNWPU

View full text Add to dashboard Cite

In multisensor systems, the signal processing delay, measurement acquisition delay, and other factors will lead to imprecisely time-stamped measurements, namely, the problem of time-offset. To deal with the measurement time offsets in distributed multisensor systems, a distributed multisensor multitarget tracking algorithm with time-offset registration is proposed. The local processors track multiple targets in the presence of false alarms and missed detections based on the joint probabilistic data association (JPDA) algorithm and the extended Kalman filter (EKF), providing the time-biased local tracks. In the global processor, in allusion to the global track accuracy degradation introduced by the time offsets of local tracks, the equivalent measurements are firstly constructed based on local tracks by using the inverse Kalman filter. The pseudo-measurement equation of time offset for constant velocity targets is derived and the pseudo-measurement calculation method is presented. Then, the pseudo-measurement based relative time-offset estimation algorithm is presented, by using the recursive least squares estimation (RLSE) and the Kalman filter (KF) to jointly estimate the state in space and time domains, respectively. Finally, a framework of distributed multisensor multitarget tracking with time-offset registration is presented, where the time-varying relative time-offset estimation and compensation, 'equivalent measurement to global track' association, and global track update are included. Simulations for multisensor multitarget tracking in the presence of false alarms and missed detections are conducted, demonstrating that the present algorithm effectively improves the accuracy of fused global tracks.

show abstract

Multi-sensor multi-target tracking with generalized labeled multi-Bernoulli filter based on track-before-detect observation model using Gaussian belief propagation

Cao,

Zhao

2024

Digital Signal Processing

View full text Add to dashboard Cite

Multi-Scan Multi-Sensor Multi-Object State Estimation

Cited by 19 publications

References 29 publications

Linear Complexity Gibbs Sampling for Generalized Labeled Multi-Bernoulli Filtering

Linear Complexity Gibbs Sampling for Generalized Labeled Multi-Bernoulli Filtering

Distributed Multisensor Multitarget Tracking Algorithm with Time-Offset Registration

Multi-sensor multi-target tracking with generalized labeled multi-Bernoulli filter based on track-before-detect observation model using Gaussian belief propagation

Contact Info

Product

Resources

About