Juan Pablo Fernández-Gutiérrez
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Juan Pablo Fernández-Gutiérrez has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- José Perea-Arango
Empresas Publicas de Medellin
1 shared publication - José-Fernando Camacho-Vallejo
Tecnológico de Monterrey, Escuela de Ingeniería y Ciencias
1 shared publication - Juan Guillermo Villegas
Universidad de Antioquía
1 shared publication Juan Pablo
1 shared publication- Nelson Eduardo Castaño Giraldo
Tecnológico de Antioquia ‐ University Institution Medellín Colombia
1 shared publication Piotr Graczyk
LAREMA Université d'Angers Angers France
1 shared publication
Latest Publications
- 23 December 2025
- Published by Wiley in Asian Journal of Control
A centralized normal barrier approach for min‐sup robust Kalman filtering with unknown covariance bounds
The min‐sup robust Kalman filter (RKF) provides robust state estimation for uncertain linear dynamic systems, relying on centered distribution classes with bounded covariances. Identifying the upper‐bound matrices that define these covariances is crucial for RKF implementation. Existing approaches, such as the penalized uncoverage function introduced in 2022, may produce overly conservative bounds. This paper introduces a new centralized normal barrier identification (CNBI) function and a pruned variant (PCNBI) to improve upper‐bound covariance matrix identification. The PCNBI methodology is applied to robust state estimation in a two‐dimensional linear system. Simulation results show that PCNBI significantly outperforms recent robust approaches, including those based on the Huber loss function, Student's distribution, Kullback–Leibler divergence, and the robust autocovariance least‐squares (RALS, 2022) method, in the presence of outliers and Laplace‐distributed noise, while achieving competitive performance with RALS under ‐contamination distributions.
- 23 December 2025
- Published by Wiley in Asian Journal of Control
A centralized normal barrier approach for min‐sup robust Kalman filtering with unknown covariance bounds
The min‐sup robust Kalman filter (RKF) provides robust state estimation for uncertain linear dynamic systems, relying on centered distribution classes with bounded covariances. Identifying the upper‐bound matrices that define these covariances is crucial for RKF implementation. Existing approaches, such as the penalized uncoverage function introduced in 2022, may produce overly conservative bounds. This paper introduces a new centralized normal barrier identification (CNBI) function and a pruned variant (PCNBI) to improve upper‐bound covariance matrix identification. The PCNBI methodology is applied to robust state estimation in a two‐dimensional linear system. Simulation results show that PCNBI significantly outperforms recent robust approaches, including those based on the Huber loss function, Student's distribution, Kullback–Leibler divergence, and the robust autocovariance least‐squares (RALS, 2022) method, in the presence of outliers and Laplace‐distributed noise, while achieving competitive performance with RALS under ‐contamination distributions.
- 4 October 2025
- Published by Springer Nature in Computational Optimization and Applications
- Vol. 93 (2) , 851-897
A simple and effective exact method for the medianoid problem with multipurpose trips
- 4 October 2025
- Published by Springer Nature in Computational Optimization and Applications
- Vol. 93 (2) , 851-897
A simple and effective exact method for the medianoid problem with multipurpose trips
- Open Access
- 1 January 2024
- Published by Universidad de Antioquia in Revista Facultad de Ingeniería Universidad de Antioquia
Minimax Regret filter for uncertainty Single-Input Single-Output systems: simulation study
The Kalman filter, widely used since its introduction in 1960, assumes Gaussian random disturbances. However, this assumption can be inappropriate in non-Gaussian contexts, leading to suboptimal performance. Researchers have proposed robust filters like minimax filters to address this limitation, but these filters can overly conservative estimates. This research introduces a novel approach that combines unknown-but-bounded dynamics for the state process and stochastic processes for the measurement equation along with a Minimax Regret framework to improve state estimation in one-dimensional linear dynamic models. We evaluate the proposed method through two simulation studies. The first study optimizes the hyperparameter value using Grid Search. In contrast, the second compares the performance of the proposed method with conventional methods, including the Kalman filter and a robust version of the RobKF filter implemented in R software, using a suitable performance metric such as mean squared error. The results demonstrate the superiority of the proposed algorithm.
- 1 January 2024
- Published by Universidad de Antioquia in Revista Facultad de Ingeniería Universidad de Antioquia
Open AccessMinimax Regret filter for uncertainty Single-Input Single-Output systems: simulation study
The Kalman filter, widely used since its introduction in 1960, assumes Gaussian random disturbances. However, this assumption can be inappropriate in non-Gaussian contexts, leading to suboptimal performance. Researchers have proposed robust filters like minimax filters to address this limitation, but these filters can overly conservative estimates. This research introduces a novel approach that combines unknown-but-bounded dynamics for the state process and stochastic processes for the measurement equation along with a Minimax Regret framework to improve state estimation in one-dimensional linear dynamic models. We evaluate the proposed method through two simulation studies. The first study optimizes the hyperparameter value using Grid Search. In contrast, the second compares the performance of the proposed method with conventional methods, including the Kalman filter and a robust version of the RobKF filter implemented in R software, using a suitable performance metric such as mean squared error. The results demonstrate the superiority of the proposed algorithm.