CS Paper Radar

Fixed-structure Gaussian Mixture Filtering with Robust Measurement Updates under Outliers

Ondřej Straka, Uwe D. Hanebeck

eess.SY ·submitted 2026-08-20 11:03 UTC · v2
published

Abstract

Bayesian state estimation for discrete-time nonlinear stochastic systems is considered in the presence of measurement outliers. Building on a fixed-structure Gaussian mixture filtering framework, this paper proposes a robust measurement-update variant in which the predictive density structure is determined by an offline decomposition of the transition density into axis-aligned Gaussian components. This construction maintains the Gaussian mixture structure as deterministic and tunable via the chosen decomposition fidelity. Measurement components affected by outliers are modeled using a Student's-t distribution, and the corresponding update of each Gaussian mixture component is approximated by a variational Bayes procedure. The resulting filter is evaluated in a three-dimensional tracking scenario with range and bearing measurements, where the bearing channel is affected by outliers, modeled as heavy-tailed noise.

Comment
Accepted for the conference IEEE MFI 2026

Closest papers

Nearest neighbours by meaning across the whole index.

Variable Dimension IMM and GPB2 Filters for Tracking in Multiple Model Systems

eess.SP ·2026-09-25 ·Roberto Pérez-Pérez, Ángel F. García-Fernández
top venuecode/data similarity 0.56
Abstract · Matlab code of VD-IMM and VD-GPB2 is provided at https://github.com/ropperez/VD_Multiple_Models

This paper presents mathematically principled filters for multiple-model systems with states of different dimensionality, specifically the variable-dimension interacting multiple model (VD-IMM) filter and the variable-dimension generalised pseudo-Bayesian filter of order 2 (VD-GPB2). To do so, we first provide a Bayesian modelling of a variable dimensional dynamic system, and its measurements. Then, for variable dimensional linear Gaussian dynamic and measurement models, the VD-IMM filter is derived by assuming a posterior that has a Gaussian density for each mode, and then performing a Kullback-Leibler Divergence (KLD) minimisation after each prediction step to keep the Gaussian density form for each mode. Subsequently, the Bayesian update step is performed, keeping a Gaussian density for each mode. The VD-GPB2 filter considers the same model as the VD-IMM filter and also assumes that the posterior for each mode is Gaussian. In contrast, the VD-GPB2 filter propagates a Gaussian mixture for each mode in the prediction step. In the update step, the VD-GPB2 filter performs a KLD minimisation to have a Gaussian density for each mode. Simulation results show the benefits of the VD-IMM and VD-GPB2 filters compared to previous alternatives.

0.55

Variational Outlier-Robust Gaussian Process Regression with Generative Modeling

cs.LG ·2026-08-17 ·Arslan Majal, Aamir Hussain Chughtai
similarity 0.56
Abstract · 5 pages, 1 figure

Outliers can substantially distort Gaussian process regression (GPR) due to its conventional Gaussian observation likelihood, leading to inaccurate model learning and prediction. To address this limitation, this article introduces a generative GPR model that captures observation-specific contamination and adaptively mitigates the influence of outliers. Subsequently, a variational generalized expectation-maximization procedure is used to learn the latent variables and GPR model parameters. Experiments on synthetic and real datasets under different contamination settings demonstrate that the proposed method remains competitive with-and in several cases outperforms-robust GPR baselines in prediction accuracy. Moreover, the proposed method shares the cubic computational scaling of the compared GPR methods.

0.23

The Zonotopic Mixture Filter

eess.SY eess.SP ·2026-08-18 ·Rodrigo A. González, Angel L. Cedeño, Vicenç Puig
similarity 0.54
Abstract · 16 pages, 8 figures

State estimation is commonly posed in either a probabilistic or an unknown-but-bounded framework. The former requires a fully specified noise distribution, typically with unbounded support, while the latter yields guaranteed enclosures that carry no probabilistic weighting. Bridging these noise descriptions, this paper proposes a zonotopic mixture noise model, in which the noise is generated by drawing a zonotope from a finite collection according to fixed probabilities and then realizing an arbitrary element of it. For this noise model, we derive the zonotopic mixture filter, which propagates a bank of zonotopic Kalman filters over mode histories, discards the histories falsified by the data, and weights the surviving ones by their relative probability. The resulting state enclosures yield guaranteed coverage probabilities and remain valid for every noise realization compatible with the bounds, and a greedy mixture reduction scheme preserves these statistical guarantees while keeping the representation tractable. Numerical examples illustrate the proposed approach and its potential benefits over related state estimation methods.

0.29

Adaptive DOA Estimation With KL-Based Kernel Bandwidth Adaptation Under Impulsive Noise

eess.SP ·2026-10-10 ·Mojtaba Hajiabadi, Saeid Pakravan, Ghosheh Abed Hodtani, Ming Zeng
similarity 0.51
Abstract

Direction-of-arrival (DOA) estimation can degrade markedly in impulsive noise, where methods relying on Gaussian or second-order statistical assumptions are particularly susceptible to large-amplitude outliers. This paper develops an adaptive DOA estimator that employs Gaussian kernel weighting to attenuate the influence of large prediction errors while adjusting the kernel bandwidth online according to a Kullback--Leibler (KL) divergence criterion. A sample-based approximation of the KL criterion yields a recursive bandwidth update that adapts the degree of outlier suppression to the observed error statistics rather than prescribing a fixed bandwidth a priori. The spatial-filter coefficients and kernel bandwidth are updated recursively, and the DOAs are recovered from the resulting spatial spectrum. Mean and mean-square analyses are developed under standard adaptive-filtering approximations to characterize the coefficient convergence and steady-state behavior. Simulation results under Gaussian-mixture impulsive noise demonstrate stable coefficient adaptation and improved DOA estimation accuracy over representative competing methods in several challenging operating regimes, with the most pronounced gains occurring at low-to-moderate signal-to-noise ratios. The results also indicate reduced sensitivity to the initial kernel bandwidth through online bandwidth adaptation.

0.23

Personalized Federated Hierarchical Gaussian Processes for Privacy-Preserving Modeling of Heterogeneous Distributed Systems

cs.LG ·2026-09-16 ·Xianjian Xie, Hao Yan
similarity 0.51
Abstract

We present Personalized Federated Hierarchical Gaussian Processes (pFedHGP) for probabilistic regression and classification when data are distributed across heterogeneous clients. Each client's latent function decomposes into (i) a shared global component, (ii) a client-specific deviation that shares the global kernel structure, and (iii) a flexible local residual. Sparse inducing-variable approximations and federated variational inference keep raw data local while the server synchronizes only low-dimensional statistics for the shared component. Full predictive distributions support uncertainty-aware decisions. In application studies, pFedHGP attains perfect fault classification in press tonnage monitoring using 13.77% of labeled cycles and recovers geographic zones in federated air-quality modeling without centralizing station-level time series. An Instantaneous Linear Mixing Model viewpoint links the hierarchy to multi-output Gaussian processes for correlated sensors.

0.33

Multi-Target Tracking of Cooperating Swarms by an Ensemble Gaussian Mixture Filter

eess.SY stat.APmath.PR ·2026-10-02 ·Zain Jabbar, Andrey A. Popov
theory similarity 0.51
Abstract

Per-target labeled random-finite-set filters, such as the state of the art $δ$-generalized labeled multi-Bernoulli filter, cannot represent cooperating, swarm-like, motion between targets. In this work, we present a new filter that can. The new particle filter couples the joint particle filter with ensemble Gaussian mixture filter such that a measurement updates the density at every particle instead of merely reweighting. We prove that the new particle filter converges to the joint particle filter in the limit of particle number and thus to the true Bayesian posterior, under certain assumptions. Experimental results on two coupled systems, the Vicsek model and coupled Brownian motion provide evidence on the efficacy of the filter.

0.18

Ensemble generative filtering for sequential data assimilation in dynamical systems

cs.CE ·2026-09-12 ·Xu-Hui Zhou, Jiequn Han
similarity 0.50
Abstract

Sequential data assimilation (DA) faces a fundamental trade-off: particle filters capture non-Gaussian cycling priors but require prohibitively large ensembles, whereas the ensemble Kalman filter (EnKF) is computationally efficient but constrained by its Gaussian assumption. As machine learning enables rapid model forecasts, exploiting non-Gaussian prior features via moderately large ensembles has become increasingly viable. To exploit this opportunity, we propose the ensemble generative filter (EnGF), a simple yet effective method for non-Gaussian filtering. The key idea is to fit a generative model to the forecast ensemble at each DA cycle and harness its defining strength, inexpensive sampling, to draw a much larger particle population for Bayesian analysis without any additional model forecasts; we adopt a Gaussian mixture model as a lightweight instance that can be fit cheaply from a moderate ensemble. To address practical challenges, we further extend the EnGF by introducing (i) a tempered EnGF using likelihood tempering to prevent particle degeneracy under informative observations and (ii) a latent EnGF that performs prior modeling and Bayesian updates in a reduced latent space for high-dimensional systems. Across chaotic systems (doubling map, Lorenz-63, and Lorenz-96) and a challenging shock-tube problem, the EnGF delivers clear and often substantial improvements over the EnKF, in some cases even approaching the filtering accuracy of a massive-ensemble particle filter at a small fraction of its cost.

0.34

Bayesian Filtering in Physical Systems via Test-time Trained Flow Matching

cs.AI stat.MLcs.LG ·2026-09-20 ·Ruiqi Feng, Chongyi Wang, Tao Zhang, Tailin Wu
benchmark similarity 0.49
Abstract

Bayesian filtering provides a principled framework for online state estimation under uncertainty, yet its application to systems with high-dimensional states and complicated posterior distributions remains challenging. Recent generative models, such as flow matching, have shown potential in Bayesian filtering. However, they still rely on particle-based representations of the posterior, which lose the rich information of the full distribution, or tackle a trajectory-level inverse problem that conflicts with the recursive structure of Bayesian filtering. To address this, we propose a new perspective of directly encoding the evolving distribution into flow matching model weights, namely, the Belief Flow Filter (BFF). It is a generative filtering framework that updates model weights via gradient descent at test time to track the posterior evolution. Thereby, BFF bypasses the scalability issue of particle representations or the flexibility limitation of Gaussian assumptions in conventional filters. We theoretically justify that the BFF design is structurally aligned with Bayesian filtering, and its training objective targets the recursive filtering operator. BFF is empirically verified across 5 different physical systems, including ones with chaotic dynamics and highly sparse, non-linear observations. The results show that BFF attains the best score in 8 of 9 metric-benchmark cells across the three standard 1D and 2D PDE benchmarks, and similarly leads on the extreme single-moving-sensor setting and on a real-world-grounded tokamak plasma estimation task, demonstrating its potential to accurately approximate the Bayesian filtering operator in high-dimensional probability space.

0.23