Abstract:The accuracy of moving horizon estimation (MHE) suffers significantly in the presence of measurement outliers. Existing methods address this issue by treating measurements leading to large MHE cost function values as outliers, which are subsequently discarded. This strategy, achieved through solving combinatorial optimization problems, is confined to linear systems to guarantee computational tractability and stability. Contrasting these heuristic solutions, our work reexamines MHE from a Bayesian perspective, unveils the fundamental issue of its lack of robustness: MHE's sensitivity to outliers results from its reliance on the Kullback-Leibler (KL) divergence, where both outliers and inliers are equally considered. To tackle this problem, we propose a robust Bayesian inference framework for MHE, integrating a robust divergence measure to reduce the impact of outliers. In particular, the proposed approach prioritizes the fitting of uncontaminated data and lowers the weight of contaminated ones, instead of directly discarding all potentially contaminated measurements, which may lead to undesirable removal of uncontaminated data. A tuning parameter is incorporated into the framework to adjust the robustness degree to outliers. Notably, the classical MHE can be interpreted as a special case of the proposed approach as the parameter converges to zero. In addition, our method involves only minor modification to the classical MHE stage cost, thus avoiding the high computational complexity associated with previous outlier-robust methods and inherently suitable for nonlinear systems. Most importantly, our method provides robustness and stability guarantees, which are often missing in other outlier-robust Bayes filters. The effectiveness of the proposed method is demonstrated on simulations subject to outliers following different distributions, as well as on physical experiment data.

Robust M-estimation Using the Equivalent Weights Constructed by Removing the Influence of an Outlier on the Residuals

Robust Kernel Recursive Adaptive Filtering Algorithms Based on M-estimate

Robust Estimation of Measurement Error Covariance

A NEW PERSPECTIVE ON ROBUST M-ESTIMATION: FINITE SAMPLE THEORY AND APPLICATIONS TO DEPENDENCE-ADJUSTED MULTIPLE TESTING

Robust explicit estimators using the power-weighted repeated medians

Robust M-Estimation for Additive Single-Index Cointegrating Time Series Models

A redescending M-estimator approach for outlier-resilient modeling

Novel Robust Least-Squares Estimator for Linear Dynamic Data Reconciliation

DETECTION OF MULTIPLE OUTLIERS BY RANDOM ROBUST TESTING

A hard re-descending hybrid robust regression estimation technique using direct weights

Double robust semiparametric weighted M-estimators of a zero-inflated Poisson regression with missing data in covariates

A New Sparse and Robust Adaptive Lasso Estimator for the Independent Contamination Model

Robust Bayesian Inference for Moving Horizon Estimation

Robust Adaptive Filtering based on M-estimation-based Minimum Error Entropy Criterion

A Robust Learning Algorithm for Regression Models Using Distributionally Robust Optimization under the Wasserstein Metric

M robust bias-corrected weighted least squares iteration solution of mixed additive and multiplicative random error model

A Robust Learning Approach for Regression Models Based on Distributionally Robust Optimization.

Improving the Robustness of the Theil-Sen Estimator Using a Simple Heuristic-Based Modification

Large Dimensional Analysis of Robust M-Estimators of Covariance with Outliers

A Robust Adaptive Modified Maximum Likelihood Estimator for the Linear Regression Model

Robust small area estimation for unit level model with density power divergence