Publications
81.
Roy, Wim Van; Abbasi-Esfeden, Ramin; Swevers, Jan
Online Unit Commitment Problem Solving using Extended Dynamic Programming Presentation
22.03.2023.
@misc{vanRoy2023EDP,
title = {Online Unit Commitment Problem Solving using Extended Dynamic Programming},
author = {Wim Van Roy and Ramin Abbasi-Esfeden and Jan Swevers},
year = {2023},
date = {2023-03-22},
urldate = {2023-03-22},
keywords = {},
pubstate = {published},
tppubtype = {presentation}
}
82.
Zhang, Shuhao; Vandewal, Bastiaan; Bos, Mathis; Decré, Wilm; Swevers, Jan
Vision-based localization and parking space detection for the truck-trailer Autonomous Mobile Robot Presentation
21.03.2023, (Abstract at the 2023 Benelux Meeting ).
@misc{lirias4066797,
title = {Vision-based localization and parking space detection for the truck-trailer Autonomous Mobile Robot},
author = {Shuhao Zhang and Bastiaan Vandewal and Mathis Bos and Wilm Decré and Jan Swevers},
year = {2023},
date = {2023-03-21},
urldate = {2024-02-07},
note = {Abstract at the 2023 Benelux Meeting },
keywords = {},
pubstate = {published},
tppubtype = {presentation}
}
83.
Bonassi, Fabio
2023.
@phdthesis{bonassi2023reconciling,
title = {Reconciling deep learning and control theory: recurrent neural networks for model-based control design},
author = {Fabio Bonassi},
url = {https://www.politesi.polimi.it/handle/10589/196384},
year = {2023},
date = {2023-02-01},
urldate = {2023-02-01},
address = {Milan, Italy},
institution = {Politecnico di Milano},
abstract = {This doctoral thesis aims to establish a theoretically-sound framework for the adoption of Recurrent Neural Network (RNN) models in the context of nonlinear system identification and model-based control design. The idea, long advocated by practitioners, of exploiting the remarkable modeling performances of RNNs to learn black-box models of unknown nonlinear systems, and then using such models to synthesize model-based control laws, has already shown considerable potential in many practical applications. On the other hand, the adoption of these architectures by the control systems community has been so far limited, mainly because the generality of these architectures makes it difficult to attain general properties and to build solid theoretical foundations for their safe and profitable use for control design. To address these gaps, we first provide a control engineer-friendly description of the most common RNN architectures, i.e., Neural NARXs (NNARXs), Gated Recurrent Units (GRUs), and Long Short-Term Memory networks (LSTMs), as well as their training procedure. The stability properties of these architectures are then analyzed, using common nonlinear systems’ stability notions such as the Input-to-State Stability (ISS), the Input-to-State Practical Stability (ISPS), and the Incremental Input-to-State Stability (δISS). In particular, sufficient conditions for these properties are devised for the considered RNN architectures, and it is shown how to enforce these conditions during the training procedure, in order to learn provenly stable RNN models. Model-based control strategies are then synthesized for these models. In particular, nonlinear model predictive control schemes are first designed: in this context, the model’s δISS is shown to enable the attainment of nominal closed-loop stability and, under a suitable design of the control scheme, also robust asymptotic zero-error output regulation. Then, an alternative computationally-lightweight control scheme, based on the internal model control strategy, is proposed, and its closed-loop properties are discussed. The performances of these control schemes are tested on several nonlinear benchmark systems, demonstrating the potentiality of the proposed framework. Finally, some fundamental issues for the practical implementation of RNN-based control strategies are mentioned. In particular, we discuss the need for the safety verification of RNN models and their adaptation in front of changes of the plant’s behavior, the definition of RNN structures that exploit qualitative physical knowledge of the system to boost the performances and interpretability of these models, and the problem of designing control schemes that are robust to the unavoidable plant-model mismatch.},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
This doctoral thesis aims to establish a theoretically-sound framework for the adoption of Recurrent Neural Network (RNN) models in the context of nonlinear system identification and model-based control design. The idea, long advocated by practitioners, of exploiting the remarkable modeling performances of RNNs to learn black-box models of unknown nonlinear systems, and then using such models to synthesize model-based control laws, has already shown considerable potential in many practical applications. On the other hand, the adoption of these architectures by the control systems community has been so far limited, mainly because the generality of these architectures makes it difficult to attain general properties and to build solid theoretical foundations for their safe and profitable use for control design. To address these gaps, we first provide a control engineer-friendly description of the most common RNN architectures, i.e., Neural NARXs (NNARXs), Gated Recurrent Units (GRUs), and Long Short-Term Memory networks (LSTMs), as well as their training procedure. The stability properties of these architectures are then analyzed, using common nonlinear systems’ stability notions such as the Input-to-State Stability (ISS), the Input-to-State Practical Stability (ISPS), and the Incremental Input-to-State Stability (δISS). In particular, sufficient conditions for these properties are devised for the considered RNN architectures, and it is shown how to enforce these conditions during the training procedure, in order to learn provenly stable RNN models. Model-based control strategies are then synthesized for these models. In particular, nonlinear model predictive control schemes are first designed: in this context, the model’s δISS is shown to enable the attainment of nominal closed-loop stability and, under a suitable design of the control scheme, also robust asymptotic zero-error output regulation. Then, an alternative computationally-lightweight control scheme, based on the internal model control strategy, is proposed, and its closed-loop properties are discussed. The performances of these control schemes are tested on several nonlinear benchmark systems, demonstrating the potentiality of the proposed framework. Finally, some fundamental issues for the practical implementation of RNN-based control strategies are mentioned. In particular, we discuss the need for the safety verification of RNN models and their adaptation in front of changes of the plant’s behavior, the definition of RNN structures that exploit qualitative physical knowledge of the system to boost the performances and interpretability of these models, and the problem of designing control schemes that are robust to the unavoidable plant-model mismatch.
84.
Laude, Emanuel; Themelis, Andreas; Patrinos, Panagiotis
Dualities for non-Euclidean smoothness and strong convexity under the light of generalized conjugacy Working paper
2023.
@workingpaper{laudeConjugateDualitiesRelative2023,
title = {Dualities for non-Euclidean smoothness and strong convexity under the light of generalized conjugacy},
author = {Emanuel Laude and Andreas Themelis and Panagiotis Patrinos},
doi = {https://doi.org/10.48550/arXiv.2112.08886},
year = {2023},
date = {2023-01-23},
urldate = {2021-01-01},
number = {arXiv:2112.08886},
publisher = {arXiv},
keywords = {},
pubstate = {published},
tppubtype = {workingpaper}
}
85.
Ghezzi, Andrea; Hoffman, Jasper; Frey, Jonathan; Boedecker, Joschka; Diehl, Moritz
Imitation Learning from Nonlinear MPC via the Exact Q-Loss and its Gauss-Newton Approximation Proceedings Article
In: 2023 Conference on Decision and Control (CDC), pp. 4766-4771, IEEE, Singapore, Singapore, 2023, ISBN: 979-8-3503-0124-3.
@inproceedings{Ghezzi2023b,
title = {Imitation Learning from Nonlinear MPC via the Exact Q-Loss and its Gauss-Newton Approximation},
author = {Andrea Ghezzi and Jasper Hoffman and Jonathan Frey and Joschka Boedecker and Moritz Diehl},
url = {https://doi.org/10.48550/arXiv.2304.01782},
doi = {10.1109/CDC49753.2023.10383323},
isbn = {979-8-3503-0124-3},
year = {2023},
date = {2023-01-19},
urldate = {2023-01-19},
booktitle = {2023 Conference on Decision and Control (CDC)},
pages = {4766-4771},
publisher = {IEEE},
address = {Singapore, Singapore},
abstract = {This work presents a novel loss function for learning nonlinear Model Predictive Control policies via Imitation Learning. Standard approaches to Imitation Learning neglect information about the expert and generally adopt a loss function based on the distance between expert and learned controls. In this work, we present a loss based on the Q-function directly embedding the performance objectives and constraint satisfaction of the associated Optimal Control Problem (OCP). However, training a Neural Network with the Q-loss requires solving the associated OCP for each new sample. To alleviate the computational burden, we derive a second Q-loss based on the Gauss-Newton approximation of the OCP resulting in a faster training time. We validate our losses against Behavioral Cloning, the standard approach to Imitation Learning, on the control of a nonlinear system with constraints. The final results show that the Q-function-based losses significantly reduce the amount of constraint violations while achieving comparable or better closed-loop costs.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This work presents a novel loss function for learning nonlinear Model Predictive Control policies via Imitation Learning. Standard approaches to Imitation Learning neglect information about the expert and generally adopt a loss function based on the distance between expert and learned controls. In this work, we present a loss based on the Q-function directly embedding the performance objectives and constraint satisfaction of the associated Optimal Control Problem (OCP). However, training a Neural Network with the Q-loss requires solving the associated OCP for each new sample. To alleviate the computational burden, we derive a second Q-loss based on the Gauss-Newton approximation of the OCP resulting in a faster training time. We validate our losses against Behavioral Cloning, the standard approach to Imitation Learning, on the control of a nonlinear system with constraints. The final results show that the Q-function-based losses significantly reduce the amount of constraint violations while achieving comparable or better closed-loop costs.
86.
Baumgärtner, Katrin; Wang, Yizhen; Zanelli, Andrea; Diehl, Moritz
Fast Nonlinear Model Predictive Control using Barrier Formulations and Squashing with a Generalized Gauss-Newton Hessian Proceedings Article
In: 2022 IEEE 61st Conference on Decision and Control (CDC), pp. 558-563, IEEE, 2023, ISBN: 978-1-6654-6761-2.
@inproceedings{Baumgaertner2022a,
title = {Fast Nonlinear Model Predictive Control using Barrier Formulations and Squashing with a Generalized Gauss-Newton Hessian},
author = {Katrin Baumgärtner and Yizhen Wang and Andrea Zanelli and Moritz Diehl},
url = {https://ieeexplore.ieee.org/document/9992869},
doi = {https://doi.org/10.1109/CDC51059.2022.9992869},
isbn = {978-1-6654-6761-2},
year = {2023},
date = {2023-01-10},
urldate = {2023-01-10},
booktitle = {2022 IEEE 61st Conference on Decision and Control (CDC)},
pages = {558-563},
publisher = {IEEE},
abstract = {We propose an approximate algorithm for Nonlinear Model Predictive Control (NMPC) which is based on a reformulation of the inequality constrained optimal control problem using barrier terms and squashing functions. Within an SQP framework, the particular structure of the reformulated problem can be leveraged by using a Generalized Gauss-Newton Hessian approximation. Moreover, the quadratic subproblems can be efficiently solved using a single backward and forward sweep of the Riccati recursion. We show local linear convergence of the proposed algorithm, as well as local quadratic convergence in the case of linear system dynamics. The computational speed-up, which can be achieved with the proposed method, is illustrated in a simulation study.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We propose an approximate algorithm for Nonlinear Model Predictive Control (NMPC) which is based on a reformulation of the inequality constrained optimal control problem using barrier terms and squashing functions. Within an SQP framework, the particular structure of the reformulated problem can be leveraged by using a Generalized Gauss-Newton Hessian approximation. Moreover, the quadratic subproblems can be efficiently solved using a single backward and forward sweep of the Riccati recursion. We show local linear convergence of the proposed algorithm, as well as local quadratic convergence in the case of linear system dynamics. The computational speed-up, which can be achieved with the proposed method, is illustrated in a simulation study.
87.
Baumgärtner, Katrin; Reiter, Rudolf; Diehl, Moritz
Moving Horizon Estimation with Adaptive Regularization for Ill-Posed State and Parameter Estimation Problems Proceedings Article
In: 2022 IEEE 61st Conference on Decision and Control (CDC), pp. 2165-2171, IEEE, Cancun, Mexico, 2023, ISBN: 978-1-6654-6761-2.
@inproceedings{Baumgärtner2022MHE,
title = {Moving Horizon Estimation with Adaptive Regularization for Ill-Posed State and Parameter Estimation Problems},
author = {Katrin Baumgärtner and Rudolf Reiter and Moritz Diehl},
doi = {10.1109/CDC51059.2022.9993416},
isbn = {978-1-6654-6761-2},
year = {2023},
date = {2023-01-10},
urldate = {2023-01-10},
booktitle = {2022 IEEE 61st Conference on Decision and Control (CDC)},
pages = {2165-2171},
publisher = {IEEE},
address = {Cancun, Mexico},
abstract = {We investigate the usage of Moving Horizon Estimation (MHE) for state and parameter estimation for partially non-detectable systems with measurements corrupted by outliers. We propose an arrival cost update formula based on the Generalized Gauss-Newton method and illustrate how it can be generalized to nonconvex loss functions that can be effectively used for outlier rejection. Moreover, we propose an adaptive regularization scheme for the arrival cost which introduces forgetting as well as additional pseudo-measurements to the arrival cost update. We illustrate the performance of the proposed algorithms on a longitudinal vehicle state and parameter estimation problem.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We investigate the usage of Moving Horizon Estimation (MHE) for state and parameter estimation for partially non-detectable systems with measurements corrupted by outliers. We propose an arrival cost update formula based on the Generalized Gauss-Newton method and illustrate how it can be generalized to nonconvex loss functions that can be effectively used for outlier rejection. Moreover, we propose an adaptive regularization scheme for the arrival cost which introduces forgetting as well as additional pseudo-measurements to the arrival cost update. We illustrate the performance of the proposed algorithms on a longitudinal vehicle state and parameter estimation problem.
88.
Messerer, Florian; Baumgärtner, Katrin; Diehl, Moritz
A Dual-Control Effect Preserving Formulation for Nonlinear Output-Feedback Stochastic Model Predictive Control With Constraints Journal Article
In: IEEE Control Systems Letters, vol. 7, no. 1171--1176, 2023.
@article{Messerer2023,
title = {A Dual-Control Effect Preserving Formulation for Nonlinear Output-Feedback Stochastic Model Predictive Control With Constraints},
author = {Florian Messerer and Katrin Baumgärtner and Moritz Diehl},
url = {https://ieeexplore.ieee.org/document/9993720},
doi = {10.1109/LCSYS.2022.3230552},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {IEEE Control Systems Letters},
volume = {7},
number = {1171--1176},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
89.
Cecchin, Leonardo; Baumgärtner, Katrin; Gering, Stefan; Diehl, Moritz
Locally Weighted Regression with Approximate Derivatives for Data-based optimization Proceedings Article
In: 2023 European Control Conference (ECC), pp. 1–6, IEEE 2023.
@inproceedings{cecchin2023locally,
title = {Locally Weighted Regression with Approximate Derivatives for Data-based optimization},
author = { Leonardo Cecchin and Katrin Baumgärtner and Stefan Gering and Moritz Diehl},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {2023 European Control Conference (ECC)},
pages = {1–6},
organization = {IEEE},
abstract = {Interpolation and approximation of data provided in terms of a Look-Up Table (LUT) is a common and well-known task, and is especially relevant for industrial applications. When using the function for point-wise evaluation, the method choice only affects the accuracy of the function value itself. However, when the LUT is used as part of an optimization problem formulation, a bad method choice can prevent convergence or alter significantly the outcome of the solver. Moreover, computational efficiency becomes critical due to the much higher number of evaluations required. This work focuses on a variation of Locally Weighted Regression, with approximate derivatives computation. The result is a method that allows one to obtain the function value together with the first n derivatives, at a reduced computational cost. Theoretical properties of the approach are analyzed, and the results of a minimization problem using the proposed method are compared with more traditional ones. The new approach shows promising performance and results, both for computational efficiency and effectiveness when used in optimization.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Interpolation and approximation of data provided in terms of a Look-Up Table (LUT) is a common and well-known task, and is especially relevant for industrial applications. When using the function for point-wise evaluation, the method choice only affects the accuracy of the function value itself. However, when the LUT is used as part of an optimization problem formulation, a bad method choice can prevent convergence or alter significantly the outcome of the solver. Moreover, computational efficiency becomes critical due to the much higher number of evaluations required. This work focuses on a variation of Locally Weighted Regression, with approximate derivatives computation. The result is a method that allows one to obtain the function value together with the first n derivatives, at a reduced computational cost. Theoretical properties of the approach are analyzed, and the results of a minimization problem using the proposed method are compared with more traditional ones. The new approach shows promising performance and results, both for computational efficiency and effectiveness when used in optimization.
90.
Cecchin, Leonardo; Frey, Jonathan; Gering, Stefan; Manderla, Maximilian; Trachte, Adrian; Diehl, Moritz
Nonlinear Model Predictive Control for Efficient Control of Variable Speed Variable Displacement Pumps Proceedings Article
In: 2023 Modeling, Estimation and Control Conference (MECC), pp. 1–6, IFAC 2023.
@inproceedings{cecchin2023nonlinear,
title = {Nonlinear Model Predictive Control for Efficient Control of Variable Speed Variable Displacement Pumps},
author = { Leonardo Cecchin and Jonathan Frey and Stefan Gering and Maximilian Manderla and Adrian Trachte and Moritz Diehl},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {2023 Modeling, Estimation and Control Conference (MECC)},
pages = {1–6},
organization = {IFAC},
abstract = {Hydraulic pumps are a key component in manufacturing industry and off-highway vehicles.
Paired with diesel engines or electric motors, they provide hydraulic flow that can conveniently be used to power a variety of actuators.
Hydraulic power transmission has numerous advantages, unfortunately energy efficiency is usually not one of those.
The use of Variable Speed Variable Displacement pumps has been proven to be advantageous with respect to constant speed or constant displacement solutions: It allows to achieve higher efficiency and faster flow tracking dynamics.
This paper presents the development of a Model Predictive Control for this system, considering the nonlinearities and look-up-tables that characterize the system dynamics.
The Model Predictive Controller is then compared both in simulation and on test bench with a reference controller for such system, showing potential both regarding efficiency and flow tracking dynamics.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Hydraulic pumps are a key component in manufacturing industry and off-highway vehicles.
Paired with diesel engines or electric motors, they provide hydraulic flow that can conveniently be used to power a variety of actuators.
Hydraulic power transmission has numerous advantages, unfortunately energy efficiency is usually not one of those.
The use of Variable Speed Variable Displacement pumps has been proven to be advantageous with respect to constant speed or constant displacement solutions: It allows to achieve higher efficiency and faster flow tracking dynamics.
This paper presents the development of a Model Predictive Control for this system, considering the nonlinearities and look-up-tables that characterize the system dynamics.
The Model Predictive Controller is then compared both in simulation and on test bench with a reference controller for such system, showing potential both regarding efficiency and flow tracking dynamics.
Paired with diesel engines or electric motors, they provide hydraulic flow that can conveniently be used to power a variety of actuators.
Hydraulic power transmission has numerous advantages, unfortunately energy efficiency is usually not one of those.
The use of Variable Speed Variable Displacement pumps has been proven to be advantageous with respect to constant speed or constant displacement solutions: It allows to achieve higher efficiency and faster flow tracking dynamics.
This paper presents the development of a Model Predictive Control for this system, considering the nonlinearities and look-up-tables that characterize the system dynamics.
The Model Predictive Controller is then compared both in simulation and on test bench with a reference controller for such system, showing potential both regarding efficiency and flow tracking dynamics.
91.
Bonassi, Fabio; Farina, Marcello; Xie, Jing; Scattolini, Riccardo
An Offset-Free Nonlinear MPC scheme for systems learned by Neural NARX models Proceedings Article
In: 2022 IEEE 61st Conference on Decision and Control (CDC), pp. 2123-2128, IEEE, 2022, ISBN: 978-1-6654-6761-2.
@inproceedings{bonassi2022offset,
title = {An Offset-Free Nonlinear MPC scheme for systems learned by Neural NARX models},
author = {Fabio Bonassi and Marcello Farina and Jing Xie and Riccardo Scattolini},
url = {https://doi.org/10.1109/CDC51059.2022.9992362
http://arxiv.org/abs/2203.16290},
doi = {10.1109/CDC51059.2022.9992362},
isbn = {978-1-6654-6761-2},
year = {2022},
date = {2022-12-06},
urldate = {2022-01-01},
booktitle = {2022 IEEE 61st Conference on Decision and Control (CDC)},
journal = {arXiv preprint arXiv:2203.16290},
pages = {2123-2128},
publisher = {IEEE},
abstract = {This paper deals with the design of nonlinear MPC controllers that provide offset-free setpoint tracking for models described by Neural Nonlinear AutoRegressive eXogenous (NNARX) networks. The NNARX model is identified from input-output data collected from the plant, and can be given a state-space representation with known measurable states made by past input and output variables, so that a state observer is not required. In the training phase, the Incremental Input-to-State Stability (δISS) property can be forced when consistent with the behavior of the plant. The δISS property is then leveraged to augment the model with an explicit integral action on the output tracking error, which allows to achieve offset-free tracking capabilities to the designed control scheme. The proposed control architecture is numerically tested on a water heating system and the achieved results are compared to those scored by another popular offset-free MPC method, showing that the proposed scheme attains remarkable performances even in presence of disturbances acting on the plant.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper deals with the design of nonlinear MPC controllers that provide offset-free setpoint tracking for models described by Neural Nonlinear AutoRegressive eXogenous (NNARX) networks. The NNARX model is identified from input-output data collected from the plant, and can be given a state-space representation with known measurable states made by past input and output variables, so that a state observer is not required. In the training phase, the Incremental Input-to-State Stability (δISS) property can be forced when consistent with the behavior of the plant. The δISS property is then leveraged to augment the model with an explicit integral action on the output tracking error, which allows to achieve offset-free tracking capabilities to the designed control scheme. The proposed control architecture is numerically tested on a water heating system and the achieved results are compared to those scored by another popular offset-free MPC method, showing that the proposed scheme attains remarkable performances even in presence of disturbances acting on the plant.
92.
Baumgärtner, Katrin; Diehl, Moritz
The Extended Gauss-Newton Method for Nonconvex Loss Functions and its Application to Time-Optimal Model Predictive Control Proceedings Article
In: 2022 American Control Conference (ACC), pp. 4973-4978, IEEE, 2022, ISBN: 978-1-6654-5196-3.
@inproceedings{Baumgaertner2022,
title = {The Extended Gauss-Newton Method for Nonconvex Loss Functions and its Application to Time-Optimal Model Predictive Control},
author = {Katrin Baumgärtner and Moritz Diehl},
url = {https://ieeexplore.ieee.org/document/9867372},
doi = {https://doi.org/10.23919/ACC53348.2022.9867372},
isbn = {978-1-6654-5196-3},
year = {2022},
date = {2022-09-05},
urldate = {2022-09-05},
booktitle = {2022 American Control Conference (ACC)},
pages = {4973-4978},
publisher = {IEEE},
abstract = {We introduce the eXtended Gauss-Newton (XGN) method, which extends the Generalized Gauss-Newton (GGN) method to a class of nonconvex loss functions with a well-behaved global minimum. We show linear local convergence of the XGN method to a stationary point, as well as global convergence in case of a linear residual function and linear constraints. As one possible application, we illustrate how the considered class of nonconvex loss functions can be used to approximate time-optimal behaviour within a Model Predictive Control (MPC) framework.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce the eXtended Gauss-Newton (XGN) method, which extends the Generalized Gauss-Newton (GGN) method to a class of nonconvex loss functions with a well-behaved global minimum. We show linear local convergence of the XGN method to a stationary point, as well as global convergence in case of a linear residual function and linear constraints. As one possible application, we illustrate how the considered class of nonconvex loss functions can be used to approximate time-optimal behaviour within a Model Predictive Control (MPC) framework.
93.
Wang, Jianhong; Wang, Jinxin; Zhang, Yuan; Gu, Yunjie; Kim, Tae-Kyun
SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning Proceedings Article
In: Advances in Neural Information Processing Systems, 2022, (Accepted at NeurIPS 2022 Conference).
@inproceedings{wang2021shaq,
title = {SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning},
author = {Jianhong Wang and Jinxin Wang and Yuan Zhang and Yunjie Gu and Tae-Kyun Kim},
url = {https://openreview.net/forum?id=BjGawodFnOy
https://arxiv.org/abs/2105.15013},
year = {2022},
date = {2022-07-04},
urldate = {2021-01-01},
booktitle = {Advances in Neural Information Processing Systems},
journal = {arXiv preprint arXiv:2105.15013},
note = {Accepted at NeurIPS 2022 Conference},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
94.
Reiter, Rudolf; Messerer, Florian; Schratter, Markus; Watzenig, Daniel; Diehl, Moritz
An Inverse Optimal Control Approach for Trajectory Prediction of Autonomous Race Cars Proceedings Article
In: 2022 European Control Conference (ECC), pp. 146–153, IEEE, London, United Kingdom, 2022, ISBN: 978-3-907144-07-7.
@inproceedings{reiter_inverse_2022,
title = {An Inverse Optimal Control Approach for Trajectory Prediction of Autonomous Race Cars},
author = {Rudolf Reiter and Florian Messerer and Markus Schratter and Daniel Watzenig and Moritz Diehl},
url = {https://ieeexplore.ieee.org/document/9838100/},
doi = {10.23919/ECC55457.2022.9838100},
isbn = {978-3-907144-07-7},
year = {2022},
date = {2022-07-01},
urldate = {2022-09-09},
booktitle = {2022 European Control Conference (ECC)},
pages = {146--153},
publisher = {IEEE},
address = {London, United Kingdom},
abstract = {This paper proposes an optimization-based approach to predict trajectories of autonomous race cars. We assume that the observed trajectory is the result of an optimization problem that trades off path progress against acceleration and jerk smoothness, and which is restricted by constraints. The algorithm predicts a trajectory by solving a parameterized nonlinear program (NLP) which contains path progress and smoothness in cost terms. By observing the actual motion of a vehicle, the parameters of prediction are updated by means of solving an inverse optimal control problem that contains the parameters of the predicting NLP as optimization variables. The algorithm therefore learns to predict the observed vehicle trajectory in a least-squares relation to measurement data and to the presumed structure of the predicting NLP. This work contributes with an algorithm that allows for accurate and interpretable predictions with sparse data. The algorithm is implemented on embedded hardware in an autonomous real-world race car that is competing in the challenge Roborace and analyzed with respect to recorded data.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper proposes an optimization-based approach to predict trajectories of autonomous race cars. We assume that the observed trajectory is the result of an optimization problem that trades off path progress against acceleration and jerk smoothness, and which is restricted by constraints. The algorithm predicts a trajectory by solving a parameterized nonlinear program (NLP) which contains path progress and smoothness in cost terms. By observing the actual motion of a vehicle, the parameters of prediction are updated by means of solving an inverse optimal control problem that contains the parameters of the predicting NLP as optimization variables. The algorithm therefore learns to predict the observed vehicle trajectory in a least-squares relation to measurement data and to the presumed structure of the predicting NLP. This work contributes with an algorithm that allows for accurate and interpretable predictions with sparse data. The algorithm is implemented on embedded hardware in an autonomous real-world race car that is competing in the challenge Roborace and analyzed with respect to recorded data.
95.
Allamaa, Jean Pierre; Listov, Petr; Auweraer, Herman Van; Jones, Colin; Son, Tong Duy
Real-time Nonlinear MPC Strategy with Full Vehicle Validation for Autonomous Driving Proceedings Article
In: 2022 American Control Conference (ACC), pp. 1982-1987, IEEE, Atlanta, GA, USA, 2022, ISBN: 978-1-6654-5196-3.
@inproceedings{Allamaa2022RTNMPCStrategy,
title = {Real-time Nonlinear MPC Strategy with Full Vehicle Validation for Autonomous Driving},
author = {Jean Pierre Allamaa and Petr Listov and Herman Van Auweraer and Colin Jones and Tong Duy Son},
url = {https://ieeexplore.ieee.org/document/9867514},
doi = {10.23919/ACC53348.2022.9867514},
isbn = {978-1-6654-5196-3},
year = {2022},
date = {2022-06-09},
booktitle = {2022 American Control Conference (ACC)},
pages = {1982-1987},
publisher = {IEEE},
address = {Atlanta, GA, USA},
abstract = {In this paper, we present the development and deployment of an embedded optimal control strategy for autonomous driving applications on a Ford Focus road vehicle. Non-linear model predictive control (NMPC) is designed and deployed on a system with hard real-time constraints. We show the properties of sequential quadratic programming (SQP) optimization solvers that are suitable for driving tasks. Importantly, the designed algorithms are validated based on a standard automotive XiL development cycle: model-in-the-loop (MiL) with high fidelity vehicle dynamics, hardware-in-the-loop (HiL) with vehicle actuation and embedded platform, and full vehicle-hardware-in-the-loop (VeHiL). The autonomous driving environment contains both virtual simulation and physical proving ground tracks. NMPC algorithms and optimal control problem formulation are fine-tuned using a deployable C code via code generation compatible with the target embedded toolchains. Finally, the developed systems are applied to autonomous collision avoidance, trajectory tracking, and lane change at high speed on city/highway and low speed at a parking environment.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present the development and deployment of an embedded optimal control strategy for autonomous driving applications on a Ford Focus road vehicle. Non-linear model predictive control (NMPC) is designed and deployed on a system with hard real-time constraints. We show the properties of sequential quadratic programming (SQP) optimization solvers that are suitable for driving tasks. Importantly, the designed algorithms are validated based on a standard automotive XiL development cycle: model-in-the-loop (MiL) with high fidelity vehicle dynamics, hardware-in-the-loop (HiL) with vehicle actuation and embedded platform, and full vehicle-hardware-in-the-loop (VeHiL). The autonomous driving environment contains both virtual simulation and physical proving ground tracks. NMPC algorithms and optimal control problem formulation are fine-tuned using a deployable C code via code generation compatible with the target embedded toolchains. Finally, the developed systems are applied to autonomous collision avoidance, trajectory tracking, and lane change at high speed on city/highway and low speed at a parking environment.
96.
Bonassi, Fabio; Farina, Marcello; Xie, Jing; Scattolini, Riccardo
On Recurrent Neural Networks for learning-based control: recent results and ideas for future developments Journal Article
In: Journal of Process Control, vol. 114, pp. 92-104, 2022, ISSN: 0959-1524.
@article{bonassi2022survey,
title = {On Recurrent Neural Networks for learning-based control: recent results and ideas for future developments},
author = {Fabio Bonassi and Marcello Farina and Jing Xie and Riccardo Scattolini},
url = {https://arxiv.org/abs/2111.13557},
doi = {10.1016/j.jprocont.2022.04.011},
issn = {0959-1524},
year = {2022},
date = {2022-01-01},
journal = {Journal of Process Control},
volume = {114},
pages = {92-104},
abstract = {This paper aims to discuss and analyze the potentialities of Recurrent Neural Networks (RNN) in control design applications. The main families of RNN are considered, namely Neural Nonlinear AutoRegressive eXogenous, Echo State Networks, Long Short Term Memory, and Gated Recurrent Units. The goal is twofold. Firstly, to survey recent results concerning the training of RNN that enjoy Input-to-State Stability (ISS) and Incremental Input-to-State Stability (𝛿ISS) guarantees. Secondly, to discuss the issues that still hinder the widespread use of RNN for control, namely their robustness, verifiability, and interpretability. The former properties are related to the so-called generalization capabilities of the networks, i.e. their consistency with the underlying real plants, even in presence of unseen or perturbed input trajectories. The latter is instead related to the possibility of providing a clear formal connection between the RNN model and the plant. In this context, we illustrate how ISS and 𝛿ISS represent a significant step towards the robustness and verifiability of the RNN models, while the requirement of interpretability paves the way to the use of physics-based networks. The design of model predictive controllers with RNN as plant’s model is also briefly discussed. Lastly, some of the main topics of the paper are illustrated on a simulated chemical system.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper aims to discuss and analyze the potentialities of Recurrent Neural Networks (RNN) in control design applications. The main families of RNN are considered, namely Neural Nonlinear AutoRegressive eXogenous, Echo State Networks, Long Short Term Memory, and Gated Recurrent Units. The goal is twofold. Firstly, to survey recent results concerning the training of RNN that enjoy Input-to-State Stability (ISS) and Incremental Input-to-State Stability (𝛿ISS) guarantees. Secondly, to discuss the issues that still hinder the widespread use of RNN for control, namely their robustness, verifiability, and interpretability. The former properties are related to the so-called generalization capabilities of the networks, i.e. their consistency with the underlying real plants, even in presence of unseen or perturbed input trajectories. The latter is instead related to the possibility of providing a clear formal connection between the RNN model and the plant. In this context, we illustrate how ISS and 𝛿ISS represent a significant step towards the robustness and verifiability of the RNN models, while the requirement of interpretability paves the way to the use of physics-based networks. The design of model predictive controllers with RNN as plant’s model is also briefly discussed. Lastly, some of the main topics of the paper are illustrated on a simulated chemical system.
97.
Acerbo, Flavia Sofia; Swevers, Jan; Tuytelaars, Tinne; Son, Tong Duy
MPC-based Imitation Learning for Safe and Human-like Autonomous Driving Workshop
2022.
@workshop{acerbo2022mpc,
title = {MPC-based Imitation Learning for Safe and Human-like Autonomous Driving},
author = {Flavia Sofia Acerbo and Jan Swevers and Tinne Tuytelaars and Tong Duy Son},
doi = {https://doi.org/10.48550/arXiv.2206.12348},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
journal = {arXiv preprint arXiv:2206.12348},
abstract = {To ensure user acceptance of autonomous vehicles (AVs), control systems are being developed to mimic human drivers from demonstrations of desired driving behaviors. Imitation learning (IL) algorithms serve this purpose, but struggle to provide safety guarantees on the resulting closed-loop system trajectories. On the other hand, Model Predictive Control (MPC) can handle nonlinear systems with safety constraints, but realizing human-like driving with it requires extensive domain knowledge. This work suggests the use of a seamless combination of the two techniques to learn safe AV controllers from demonstrations of desired driving behaviors, by using MPC as a differentiable control layer within a hierarchical IL policy. With this strategy, IL is performed in closed-loop and end-to-end, through parameters in the MPC cost, model or constraints. Experimental results of this methodology are analyzed for the design of a lane keeping control system, learned via behavioral cloning from observations (BCO), given human demonstrations on a fixed-base driving simulator.},
keywords = {},
pubstate = {published},
tppubtype = {workshop}
}
To ensure user acceptance of autonomous vehicles (AVs), control systems are being developed to mimic human drivers from demonstrations of desired driving behaviors. Imitation learning (IL) algorithms serve this purpose, but struggle to provide safety guarantees on the resulting closed-loop system trajectories. On the other hand, Model Predictive Control (MPC) can handle nonlinear systems with safety constraints, but realizing human-like driving with it requires extensive domain knowledge. This work suggests the use of a seamless combination of the two techniques to learn safe AV controllers from demonstrations of desired driving behaviors, by using MPC as a differentiable control layer within a hierarchical IL policy. With this strategy, IL is performed in closed-loop and end-to-end, through parameters in the MPC cost, model or constraints. Experimental results of this methodology are analyzed for the design of a lane keeping control system, learned via behavioral cloning from observations (BCO), given human demonstrations on a fixed-base driving simulator.
98.
Saccani, Danilo; Cecchin, Leonardo; Fagiano, Lorenzo
Multitrajectory Model Predictive Control for Safe UAV Navigation in an Unknown Environment Journal Article
In: IEEE Transactions on Control Systems Technology, pp. 1-16, 2022.
@article{9938397,
title = {Multitrajectory Model Predictive Control for Safe UAV Navigation in an Unknown Environment},
author = {Danilo Saccani and Leonardo Cecchin and Lorenzo Fagiano},
doi = {10.1109/TCST.2022.3216989},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
journal = {IEEE Transactions on Control Systems Technology},
pages = {1-16},
abstract = {The problem of navigating an unmanned aerial vehicle (UAV) in an unknown environment is addressed with a novel model predictive control (MPC) formulation, named multitrajectory MPC (mt-MPC). The objective is to safely drive the vehicle to the desired target location by relying only on the partial description of the surroundings provided by an exteroceptive sensor. This information results in time-varying constraints during the navigation among obstacles. The proposed mt-MPC generates a sequence of position set points that are fed to control loops at lower hierarchical levels. To do so, the mt-MPC predicts two different state trajectories, a safe one and an exploiting one, in the same finite horizon optimal control problem (FHOCP). This formulation, particularly suitable for problems with uncertain time-varying constraints, allows one to partially decouple constraint satisfaction (safety) from cost function minimization (exploitation). Uncertainty due to modeling errors and sensors noise is taken into account as well, in a set membership (SM) framework. Theoretical guarantees of persistent obstacle avoidance are derived under suitable assumptions, and the approach is demonstrated experimentally out-of-the-laboratory on a prototype built with off-the-shelf components.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The problem of navigating an unmanned aerial vehicle (UAV) in an unknown environment is addressed with a novel model predictive control (MPC) formulation, named multitrajectory MPC (mt-MPC). The objective is to safely drive the vehicle to the desired target location by relying only on the partial description of the surroundings provided by an exteroceptive sensor. This information results in time-varying constraints during the navigation among obstacles. The proposed mt-MPC generates a sequence of position set points that are fed to control loops at lower hierarchical levels. To do so, the mt-MPC predicts two different state trajectories, a safe one and an exploiting one, in the same finite horizon optimal control problem (FHOCP). This formulation, particularly suitable for problems with uncertain time-varying constraints, allows one to partially decouple constraint satisfaction (safety) from cost function minimization (exploitation). Uncertainty due to modeling errors and sensors noise is taken into account as well, in a set membership (SM) framework. Theoretical guarantees of persistent obstacle avoidance are derived under suitable assumptions, and the approach is demonstrated experimentally out-of-the-laboratory on a prototype built with off-the-shelf components.
99.
Allamaa, Jean Pierre; Patrinos, Panagiotis; Auweraer, Herman Van; Son, Tong Duy
Sim2real for Autonomous Vehicle Control using Executable Digital Twin Proceedings Article
In: 10th IFAC Symposium on Advances in Automotive Control (AAC), pp. 385-391, Elsevier Ltd, 2022, ISSN: 2405-8963.
@inproceedings{ALLAMAA2022385,
title = {Sim2real for Autonomous Vehicle Control using Executable Digital Twin},
author = {Jean Pierre Allamaa and Panagiotis Patrinos and Herman Van Auweraer and Tong Duy Son},
url = {https://www.sciencedirect.com/science/article/pii/S2405896322023461},
doi = {https://doi.org/10.1016/j.ifacol.2022.10.314},
issn = {2405-8963},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
booktitle = {10th IFAC Symposium on Advances in Automotive Control (AAC)},
journal = {IFAC-PapersOnLine},
volume = {55},
number = {24},
pages = {385-391},
publisher = {Elsevier Ltd},
abstract = {In this work, we propose a sim2real method to transfer and adapt a nonlinear model predictive controller (NMPC) from simulation to the real target system based on executable digital twin (xDT). The xDT model is a high fidelity vehicle dynamics simulator, executable online in the control parameter randomization and learning process. The parameters are adapted to gradually improve control performance and deal with changing real-world environment. In particular, the performance metric is not required to be differentiable nor analytical with respect to the control parameters and system dynamics are not necessary linearized. Eventually, the proposed sim2real framework leverages altogether online high fidelity simulator, data-driven estimations, and simulation based optimization to transfer and adapt efficiently a controller developed in simulation environment to the real platform. Our experiment demonstrates that a high control performance is achieved without tedious time and labor consuming tuning.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this work, we propose a sim2real method to transfer and adapt a nonlinear model predictive controller (NMPC) from simulation to the real target system based on executable digital twin (xDT). The xDT model is a high fidelity vehicle dynamics simulator, executable online in the control parameter randomization and learning process. The parameters are adapted to gradually improve control performance and deal with changing real-world environment. In particular, the performance metric is not required to be differentiable nor analytical with respect to the control parameters and system dynamics are not necessary linearized. Eventually, the proposed sim2real framework leverages altogether online high fidelity simulator, data-driven estimations, and simulation based optimization to transfer and adapt efficiently a controller developed in simulation environment to the real platform. Our experiment demonstrates that a high control performance is achieved without tedious time and labor consuming tuning.
100.
Bonassi, Fabio; Xie, Jing; Farina, Marcello; Scattolini, Riccardo
Towards lifelong learning of Recurrent Neural Networks for control design Proceedings Article
In: 2022 European Control Conference (ECC), pp. 2018–2023, IEEE 2022.
@inproceedings{Bonassi2022,
title = {Towards lifelong learning of Recurrent Neural Networks for control design},
author = {Fabio Bonassi and Jing Xie and Marcello Farina and Riccardo Scattolini},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
booktitle = {2022 European Control Conference (ECC)},
pages = {2018--2023},
organization = {IEEE},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}