arXiv subjects

Simone Formentin

Publications and source records attributed to Simone Formentin.

At least 37 records · Page 2Linked to original sources

arXiv2024

On the impact of regularization in data-driven predictive control

Model predictive control (MPC) is a control strategy widely used in industrial applications. However, its implementation typically requires a mathematical model of the system being controlled, which can be a time-consuming and expensive task. Data-driven predictive control (DDPC) methods offer an alternative approach that does not require an explicit mathematical model, but instead optimize the control policy directly from data. In this paper, we study the impact of two different regularization penalties on the closed-loop performance of a recently introduced data-driven method called $γ$-DDPC. Moreover, we discuss the tuning of the related coefficients in different data and noise scenarios, to provide some guidelines for the end user.

eess.SY↗

arXiv2024

Meta-learning of data-driven controllers with automatic model reference tuning: theory and experimental case study

Data-driven control offers a viable option for control scenarios where constructing a system model is expensive or time-consuming. Nonetheless, many of these algorithms are not entirely automated, often necessitating the adjustment of multiple hyperparameters through cumbersome trial-and-error processes and demanding significant amounts of data. In this paper, we explore a meta-learning approach to leverage potentially existing prior knowledge about analogous (though not identical) systems, aiming to reduce both the experimental workload and ease the tuning of the available degrees of freedom. We validate this methodology through an experimental case study involving the tuning of proportional, integral (PI) controllers for brushless DC (BLDC) motors with variable loads and architectures.

eess.SY↗

arXiv2024

Sentiment-driven prediction of financial returns: a Bayesian-enhanced FinBERT approach

Predicting financial returns accurately poses a significant challenge due to the inherent uncertainty in financial time series data. Enhancing prediction models' performance hinges on effectively capturing both social and financial sentiment. In this study, we showcase the efficacy of leveraging sentiment information extracted from tweets using the FinBERT large language model. By meticulously curating an optimal feature set through correlation analysis and employing Bayesian-optimized Recursive Feature Elimination for automatic feature selection, we surpass existing methodologies, achieving an F1-score exceeding 70% on the test set. This success translates into demonstrably higher cumulative profits during backtested trading. Our investigation focuses on real-world SPY ETF data alongside corresponding tweets sourced from the StockTwits platform.

cs.CE↗

arXiv2024

Model Predictive Control with adaptive resilience for Denial-of-Service Attacks mitigation on a Regulated Dam

In recent years, SCADA (Supervisory Control and Data Acquisition) systems have increasingly become the target of cyber attacks. SCADAs are no longer isolated, as web-based applications expose strategic infrastructures to the outside world connection. In a cyber-warfare context, we propose a Model Predictive Control (MPC) architecture with adaptive resilience, capable of guaranteeing control performance in normal operating conditions and driving towards resilience against DoS (controller-actuator) attacks when needed. Since the attackers' goal is typically to maximize the system damage, we assume they solve an adversarial optimal control problem. An adaptive resilience factor is then designed as a function of the intensity function of a Hawkes process, a point process model estimating the occurrence of random events in time, trained on a moving window to estimate the return time of the next attack. We demonstrate the resulting MPC strategy's effectiveness in 2 attack scenarios on a real system with actual data, the regulated Olginate dam of Lake Como.

eess.SY↗

arXiv2024

Explainable data-driven modeling via mixture of experts: towards effective blending of grey and black-box models

Traditional models grounded in first principles often struggle with accuracy as the system's complexity increases. Conversely, machine learning approaches, while powerful, face challenges in interpretability and in handling physical constraints. Efforts to combine these models often often stumble upon difficulties in finding a balance between accuracy and complexity. To address these issues, we propose a comprehensive framework based on a "mixture of experts" rationale. This approach enables the data-based fusion of diverse local models, leveraging the full potential of first-principle-based priors. Our solution allows independent training of experts, drawing on techniques from both machine learning and system identification, and it supports both collaborative and competitive learning paradigms. To enhance interpretability, we penalize abrupt variations in the expert's combination. Experimental results validate the effectiveness of our approach in producing an interpretable combination of models closely resembling the target phenomena.

cs.LG↗

arXiv2023

Hawkes-based cryptocurrency forecasting via Limit Order Book data

Accurately forecasting the direction of financial returns poses a formidable challenge, given the inherent unpredictability of financial time series. The task becomes even more arduous when applied to cryptocurrency returns, given the chaotic and intricately complex nature of crypto markets. In this study, we present a novel prediction algorithm using limit order book (LOB) data rooted in the Hawkes model, a category of point processes. Coupled with a continuous output error (COE) model, our approach offers a precise forecast of return signs by leveraging predictions of future financial interactions. Capitalizing on the non-uniformly sampled structure of the original time series, our strategy surpasses benchmark models in both prediction accuracy and cumulative profit when implemented in a trading environment. The efficacy of our approach is validated through Monte Carlo simulations across 50 scenarios. The research draws on LOB measurements from a centralized cryptocurrency exchange where the stablecoin Tether is exchanged against the U.S. dollar.

q-fin.ST↗

arXiv2023

In-context learning of state estimators

State estimation has a pivotal role in several applications, including but not limited to advanced control design. Especially when dealing with nonlinear systems state estimation is a nontrivial task, often entailing approximations and challenging fine-tuning phases. In this work, we propose to overcome these challenges by formulating an in-context state-estimation problem, enabling us to learn a state estimator for a class of (nonlinear) systems abstracting from particular instances of the state seen during training. To this end, we extend an in-context learning framework recently proposed for system identification, showing via a benchmark numerical example that this approach allows us to (i) use training data directly for the design of the state estimator, (ii) not requiring extensive fine-tuning procedures, while (iii) achieving superior performance compared to state-of-the-art benchmarks.

eess.SY↗

arXiv2023

Unraveling the Control Engineer's Craft with Neural Networks

Many industrial processes require suitable controllers to meet their performance requirements. More often, a sophisticated digital twin is available, which is a highly complex model that is a virtual representation of a given physical process, whose parameters may not be properly tuned to capture the variations in the physical process. In this paper, we present a sim2real, direct data-driven controller tuning approach, where the digital twin is used to generate input-output data and suitable controllers for several perturbations in its parameters. State-of-the art neural-network architectures are then used to learn the controller tuning rule that maps input-output data onto the controller parameters, based on artificially generated data from perturbed versions of the digital twin. In this way, as far as we are aware, we tackle for the first time the problem of re-calibrating the controller by meta-learning the tuning rule directly from data, thus practically replacing the control engineer with a machine learning model. The benefits of this methodology are illustrated via numerical simulations for several choices of neural-network architectures.

eess.SY↗

arXiv2023

Split-Boost Neural Networks

The calibration and training of a neural network is a complex and time-consuming procedure that requires significant computational resources to achieve satisfactory results. Key obstacles are a large number of hyperparameters to select and the onset of overfitting in the face of a small amount of data. In this framework, we propose an innovative training strategy for feed-forward architectures - called split-boost - that improves performance and automatically includes a regularizing behaviour without modeling it explicitly. Such a novel approach ultimately allows us to avoid explicitly modeling the regularization term, decreasing the total number of hyperparameters and speeding up the tuning phase. The proposed strategy is tested on a real-world (anonymized) dataset within a benchmark medical insurance design problem.

cs.LG↗

arXiv2023

Optimization tools for Twin-in-the-Loop vehicle control design: analysis and yaw-rate tracking case study

Given the urgent need of simplifying the end-of-line tuning of complex vehicle dynamics controllers, the Twin-in-the-Loop Control (TiL-C) approach was recently proposed in the automotive field. In TiL-C, a digital twin is run on-board to compute a nominal control action in run-time and an additional block C_delta is used to compensate for the mismatch between the simulator and the real vehicle. As the digital twin is assumed to be the best replica available of the real plant, the key issue in TiL-C becomes the tuning of the compensator, which must be performed relying on data only. In this paper, we investigate the use of different black-box optimization techniques for the calibration of C_delta. More specifically, we compare the originally proposed Bayesian Optimization (BO) approach with the recently developed Set Membership Global Optimization (SMGO) and Virtual Reference Feedback Tuning (VRFT), a one-shot direct data-driven design method. The analysis will be carried out within a professional multibody simulation environment on a novel TiL-C application case study -- the yaw-rate tracking problem -- so as to further prove the TiL-C effctiveness on a challenging problem. Simulations will show that the VRFT approach is capable of providing a well tuned controller after a single iteration, while 10 to 15 iterations are necessary for refining it with global optimizers. Also, SMGO is shown to significantly reduce the computational effort required by BO.

eess.SY↗

arXiv2023

Joint vehicle state and parameters estimation via Twin-in-the-Loop observers

Vehicular control systems are required to be both extremely reliable and robust to different environmental conditions, e.g. load or tire-road friction. In this paper, we extend a new paradigm for state estimation, called Twin-in-the-Loop filtering (TiL-F), to the estimation of the unknown parameters describing the vehicle operating conditions. In such an approach, a digital-twin of the vehicle (usually already available to the car manufacturer) is employed on-board as a plant replica within a closed-loop scheme, and the observer gains are tuned purely from experimental data. The proposed approach is validated against experimental data, showing to significantly outperform the state-of-the-art solutions.

eess.SY↗

arXiv2023

Scenario-based model predictive control of water reservoir systems

The optimal operation of water reservoir systems is a challenging task involving multiple conflicting objectives. The main source of complexity is the presence of the water inflow, which acts as an exogenous, highly uncertain disturbance on the system. When model predictive control (MPC) is employed, the optimal water release is usually computed based on the (predicted) trajectory of the inflow. This choice may jeopardize the closed-loop performance when the actual inflow differs from its forecast. In this work, we consider - for the first time - a stochastic MPC approach for water reservoirs, in which the control is optimized based on a set of plausible future inflows directly generated from past data. Such a scenario-based MPC strategy allows the controller to be more cautious, counteracting droughty periods (e.g., the lake level going below the dry limit) while at the same time guaranteeing that the agricultural water demand is satisfied. The method's effectiveness is validated through extensive Monte Carlo tests using actual inflow data from Lake Como, Italy.

eess.SY↗

arXiv2023

Meta-learning for model-reference data-driven control

One-shot direct model-reference control design techniques, like the Virtual Reference Feedback Tuning (VRFT) approach, offer time-saving solutions for the calibration of fixed-structure controllers for dynamic systems. Nonetheless, such methods are known to be highly sensitive to the quality of the available data, often requiring long and costly experiments to attain acceptable closed-loop performance. These features might prevent the widespread adoption of such techniques, especially in low-data regimes. In this paper, we argue that the inherent similarity of many industrially relevant systems may come at hand, offering additional information from plants that are similar (yet not equal) to the system one aims to control. Assuming that this supplementary information is available, we propose a novel, direct design approach that leverages the data from similar plants, the knowledge of controllers calibrated on them, and the corresponding closed-loop performance to enhance model-reference control design. More specifically, by constructing the new controller as a combination of the available ones, our approach exploits all the available priors following a meta-learning philosophy, while ensuring non-decreasing performance. An extensive numerical analysis supports our claims, highlighting the effectiveness of the proposed method in achieving performance comparable to iterative approaches, while at the same time retaining the efficiency of one-shot direct data-driven methods like VRFT.

eess.SY↗

arXiv2023

Model predictive control with dynamic move blocking

Model Predictive Control (MPC) has proven to be a powerful tool for the control of systems with constraints. Nonetheless, in many applications, a major challenge arises, that is finding the optimal solution within a single sampling instant to apply a receding-horizon policy. In such cases, many suboptimal solutions have been proposed, among which the possibility of "blocking" some moves a-priori. In this paper, we propose a dynamic approach to move blocking, to exploit the solution already available at the previous iteration, and we show not only that such an approach preserves asymptotic stability, but also that the decrease of performance with respect to the ideal solution can be theoretically bounded.

eess.SY↗

arXiv2023

The Twin-in-the-Loop approach for vehicle dynamics control

In vehicle dynamics control, engineering a suitable regulator is a long and costly process. The starting point is usually the design of a nominal controller based on a simple control-oriented model and its testing on a full-fledged simulator. Then, many driving hours are required during the End-of-Line (EoL) tuning phase to calibrate the controller for the physical vehicle. Given the recent technological advances, in this paper we consider the pioneering perspective where the simulator can be run on-board in the electronic control unit, so as to calculate the nominal control action in real-time. In this way, it can be shown that, in the EoL phase, we only need to tune a simple compensator of the mismatch between the expected and the measured outputs. The resulting approach not only exploits the already available simulator and nominal controller and significantly simplifies the design process, but also outperforms the state-of-the-art in terms of tracking accuracy and robustness within a challenging active braking control case study.

eess.SY↗

arXiv2023

META-SMGO-$Δ$: similarity as a prior in black-box optimization

When solving global optimization problems in practice, one often ends up repeatedly solving problems that are similar to each others. By providing a rigorous definition of similarity, in this work we propose to incorporate the META-learning rationale into SMGO-$Δ$, a global optimization approach recently proposed in the literature, to exploit priors obtained from similar past experience to efficiently solve new (similar) problems. Through a benchmark numerical example we show the practical benefits of our META-extension of the baseline algorithm, while providing theoretical bounds on its performance.

math.OC↗

arXiv2023

Twin-in-the-loop state estimation for vehicle dynamics control: theory and experiments

In vehicle dynamics control, many variables of interest cannot be directly measured, as sensors might be costly, fragile, or even not available. Therefore, real-time estimation techniques need to be used. The previous approach suffers from two main drawbacks: (i) the approximations due to model mismatch might jeopardize the performance of the final estimation-based control; (ii) each new estimator requires the calibration from scratch of a dedicated model. In this paper, we propose a twin-in-the-loop scheme, where the ad-hoc model is replaced by an accurate multibody simulator of the vehicle, typically available to vehicles manufacturers and suitable for the estimation of any onboard variable, coupled with a compensator within a closed-loop observer scheme. Given the black-box nature of the digital twin, a data-driven methodology for observer tuning is developed, based on Bayesian optimization. The effectiveness of the proposed estimation method for the estimation of vehicle states and forces, as compared to traditional model-based Kalman filtering, is experimentally shown on a dataset collected with a sports car.

eess.SY↗

arXiv2023

Experience in Engineering Complex Systems: Active Preference Learning with Multiple Outcomes and Certainty Levels

Black-box optimization refers to the optimization problem whose objective function and/or constraint sets are either unknown, inaccessible, or non-existent. In many applications, especially with the involvement of humans, the only way to access the optimization problem is through performing physical experiments with the available outcomes being the preference of one candidate with respect to one or many others. Accordingly, the algorithm so-called Active Preference Learning has been developed to exploit this specific information in constructing a surrogate function based on standard radial basis functions, and then forming an easy-to-solve acquisition function which repetitively suggests new decision vectors to search for the optimal solution. Based on this idea, our approach aims to extend the algorithm in such a way that can exploit further information effectively, which can be obtained in reality such as: 5-point Likert type scale for the outcomes of the preference query (i.e., the preference can be described in not only "this is better than that" but also "this is much better than that" level), or multiple outcomes for a single preference query with possible additive information on how certain the outcomes are. The validation of the proposed algorithm is done through some standard benchmark functions, showing a promising improvement with respect to the state-of-the-art algorithm.

cs.LG↗