Anh‐Tu Nguyen

Centre National de la Recherche Scientifique

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Thierry‐Marie Guerra

Laboratoire d'Automatique, de Mécanique et d'Informatique Industrielles et Humaines

Chouki Sentouh

Université Polytechnique Hauts-de-France

Michel Dambrine

Centre National de la Recherche Scientifique

Jean‐Christophe Popieul

Université Polytechnique Hauts-de-France

Antoine Dequidt

Centre National de la Recherche Scientifique

Reinaldo M. Palhares

Universidade Federal de Minas Gerais

Jimmy Lauber

Laboratoire d'Automatique, de Mécanique et d'Informatique Industrielles et Humaines

Juntao Pan

North Minzu University

Erman Tjiputra

Jeff Clune

University of British Columbia

Cooperative Institutions

Centre National de la Recherche Scientifique

Laboratoire d'Automatique, de Mécanique et d'Informatique Industrielles et Humaines

Université Polytechnique Hauts-de-France

Vietnam National University Ho Chi Minh City

Hanoi University of Science and Technology

Ho Chi Minh City University of Technology

Vietnam Academy of Science and Technology

Hanoi University of Industry

INSA Hauts-de-France

Université de Lille

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Takagi-Sugeno Model-based Steering Control for Autonomous Vehicles with Actuator Saturation**This work has been done in the framework of the CoCoVeA research program, funded by the National Research Agency. This work was also sponsored by the International Campus on Safety and Intermodality in Transportation, the Nord-Pas-de-Calais Region, the European Community, the Regional Delegation for Research and Technology, the Ministry of Higher Education and Research, and the French National Center …

IFAC-PapersOnLine (2016)

Anh‐Tu Nguyen Chouki Sentouh Jean‐Christophe Popieul

This paper is devoted to the development of a novel control law for an autonomous vehicle. The control goal is to perform lane keeping under multiple constraints, namely actuator saturation of the steering system, roads with unknown curvature, and uncertain lateral wind force. Such system constraints will be explicitly considered in the control design procedure. To achieve this goal, we propose a novel constrained Takagi-Sugeno control method using non-quadratic Lyapunov stability framework. The obtained non-parallel distributed compensation controller is not only able to handle various system constraints but also a large variation range of vehicle speed. Especially, by means of the well-known Taylor's approximation method we can significantly reduce the numerical complexity of the proposed control method to ease the real-time control implementation. The control synthesis conditions are expressed in terms of linear matrix inequalities which can be efficiently solved with available solvers. The effectiveness of the proposed method is demonstrated with different simulation scenarios.

Linear matrix inequality

10.1016/j.ifacol.2016.07.114

Cite

Citations (7)

LMI-Based Stability Analysis for Piecewise Multi-affine Systems

IEEE Transactions on Fuzzy Systems (2016)

Anh‐Tu Nguyen Michio Sugeno Víctor Costa da Silva Campos Michel Dambrine

This paper provides a computational method to study the asymptotic stability of piecewise multi-affine (PMA) systems. Such systems stem from a class of fuzzy systems with singleton consequents and can be used to approximate any smooth nonlinear system with arbitrary accuracy. Based on the choice of piecewise Lyapunov functions, stability conditions are expressed as a feasibility test of a convex optimization with linear matrix inequality constraints. The basic idea behind these conditions is to exploit the parametric expressions of PMA systems by means of Finsler's lemma. Numerical examples are given to point out the effectiveness of the proposed method.

Linear matrix inequality

Lemma (botany)

10.1109/tfuzz.2016.2566798

Cite

Citations (17)

Robust Observer-Based Tracking Control Design for Power-Assisted Wheelchairs

IFAC-PapersOnLine (2019)

Guoxi Feng Thierry‐Marie Guerra Anh‐Tu Nguyen Lucian Buşoniu Sami Mohammad

Power-assisted wheelchairs (PAW) are efficient means of transportation for disabled persons. The resulting human-machine system includes several unknown parameters such as the mass of the user or ground adhesion. Moreover, the torque signals produced by the human are required to design a robust assistive strategy, but measuring them with torque sensors increases significantly the cost of the system. Therefore, we propose a robust observer-based assistive controller using a polytopic representation. The closed-loop control design is composed of two elements: a state feedback controller with the full state at the input, and an unknown input observer to estimate human torques and feed them into the obtained controller. The goal is to guarantee an imposed H∞ estimation performance while achieving reference tracking. To achieve the predefined performance, the observer gains are computed by solving an LMI problem. Finally, simulation results validate the control design. The methodology follows patent WO2015173094 issued in 2015 (Mohammad et al. 2015).

Observer (physics)

10.1016/j.ifacol.2019.09.119

Cite

Citations (2)

SAM: The Sensitivity of Attribution Methods to Hyperparameters

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) (2020)

Naman Bansal Chirag Agarwal Anh‐Tu Nguyen

Attribution methods can provide powerful insights into the reasons for a classifier's decision. We argue that a key desideratum of an explanation method is its robustness to input hyperparameters which are often randomly set or empirically tuned. High sensitivity to arbitrary hyperparameter choices does not only impede reproducibility but also questions the correctness of an explanation and impairs the trust of end-users. In this paper, we provide a thorough empirical study on the sensitivity of existing attribution methods. We found an alarming trend that many methods are highly sensitive to changes in their common hyperparameters e.g. even changing a random seed can yield a different explanation! Interestingly, such sensitivity is not reflected in the average explanation accuracy scores over the dataset as commonly reported in the literature. In addition, explanations generated for robust classifiers (i.e. which are trained to be invariant to pixel-wise perturbations) are surprisingly more robust than those generated for regular classifiers.

Hyperparameter

Robustness

10.1109/cvprw50498.2020.00009

Cite

Citations (13)

Double Trouble: How to not Explain a Text Classifier’s Decisions Using Counterfactuals Synthesized by Masked Language Models?

Thang M. Pham Trung Bui Long Mai Anh‐Tu Nguyen

Explaining how important each input feature is to a classifier's decision is critical in high-stake applications. An underlying principle behind dozens of explanation methods is to take the prediction difference between before-and-after an input feature (here, a token) is removed as its attribution - the individual treatment effect in causal inference. A recent method called Input Marginalization (IM) (Kim et al., 2020) uses BERT to replace a token - i.e. simulating the do(.) operator - yielding more plausible counterfactuals. However, our rigorous evaluation using five metrics and on three datasets found IM explanations to be consistently more biased, less accurate, and less plausible than those derived from simply deleting a word.

Counterfactual conditional

10.18653/v1/2022.aacl-main.2

Cite

Citations (2)

Meet in the Middle: A New Pre-training Paradigm

arXiv (Cornell University) (2023)

Anh‐Tu Nguyen Nikos Karampatziakis Weizhu Chen

Most language models (LMs) are trained and applied in an autoregressive left-to-right fashion, assuming that the next token only depends on the preceding ones. However, this assumption ignores the potential benefits of using the full sequence information during training, and the possibility of having context from both sides during inference. In this paper, we propose a new pre-training paradigm with techniques that jointly improve the training data efficiency and the capabilities of the LMs in the infilling task. The first is a training objective that aligns the predictions of a left-to-right LM with those of a right-to-left LM, trained on the same data but in reverse order. The second is a bidirectional inference procedure that enables both LMs to meet in the middle. We show the effectiveness of our pre-training paradigm with extensive experiments on both programming and natural language models, outperforming strong baselines.

Sequence (biology)

Training set

10.48550/arxiv.2303.07295

Cite

Citations (2)

Unknown Input Observers for Simultaneous Estimation of Vehicle Dynamics and Driver Torque: Theoretical Design and Hardware Experiments

IEEE/ASME Transactions on Mechatronics (2019)

Anh‐Tu Nguyen Thierry‐Marie Guerra Chouki Sentouh Hui Zhang

This article investigates a new observer design method to estimate simultaneously both the vehicle dynamics and the unknown driver's torque. To take into account the time-varying nature of the longitudinal speed, the vehicle system is transformed into a polytopic linear parameter-varying (LPV) model with a reduced level of numerical complexity. Based on Lyapunov stability arguments, we prove that the estimation errors of the system state and of the unknown input (UI) are norm-bounded, which can be made arbitrarily small by minimizing a guaranteed L _∞ -gain performance. The design of the LPV UI observer is reformulated as an linear matrix inequality-based optimization which can be effectively solved via semidefinite programming. Extensive hardware experiments are carried out under various driving test scenarios to confirm the effectiveness of the proposed observer design. In particular, a comparative study is performed with a widely adopted observer to emphasize the practical interests of the new estimation solution.

Semidefinite Programming

Observer (physics)

Linear matrix inequality

10.1109/tmech.2019.2933744

Cite

Citations (54)

Translating Simulation Images to X-Ray Images via Multi-scale Semantic Matching

Lecture notes in computer science (2024)

Jingxuan Kang Tudor Jianu Baoru Huang Binod Bhattarai Ngan Le

10.1007/978-3-031-73748-0_10

Cite

Citations (0)