Searcharxiv⌕ Search

arXiv subjects

Daniele Ravasio

Publications and source records attributed to Daniele Ravasio.

5 recordsLinked to original sources

Physics-informed structured learning of a class of recurrent neural networks with guaranteed properties

This paper proposes a physics-informed learning framework for a class of recurrent neural networks tailored to networked systems. The approach aims to learn control-oriented models that preserve the structural and stability properties of the plant. The learning algorithm is formulated as a convex optimisation problem, allowing the inclusion of linear matrix inequality constraints to enforce desired plant features. Furthermore, when the plant exhibits structural modularity, the resulting optimisation problem can be parallelised, requiring communication only among neighbouring subplants. Simulation results show the effectiveness of the proposed approach.

eess.SY↗

Decentralised model predictive control of recurrent neural network models

This work proposes a decentralised nonlinear model predictive control algorithm for systems described by input-coupled recurrent neural network models. The proposed approach exploits closed-loop incremental input-to-state stability to design the local controllers and provide robustness to subsystem couplings. The resulting control scheme guarantees asymptotic convergence to a desired equilibrium while satisfying input and output constraints. Simulation results illustrate the performance of the proposed approach.

eess.SY↗

Learning stabilising policies for constrained nonlinear systems

This work proposes a two-layered control scheme for constrained nonlinear systems represented by a class of recurrent neural networks and affected by additive disturbances. In particular, a base controller ensures global or regional closed-loop l_p-stability of the error in tracking a desired equilibrium and the satisfaction of input and output constraints within a robustly positive invariant set. An additional control contribution, derived by combining the internal model control principle with a stable operator, is introduced to improve system performance. This operator, implemented as a stable neural network, can be trained via unconstrained optimisation on a chosen performance metric, without compromising closed-loop equilibrium tracking or constraint satisfaction, even if the optimisation is stopped prematurely. In addition, we characterise the class of closed-loop stable behaviours that can be achieved with the proposed architecture. Simulation results on a pH-neutralisation benchmark demonstrate the effectiveness of the proposed approach.

eess.SY↗

Recurrent neural network-based robust control systems with regional properties and application to MPC design

This paper investigates the design of output-feedback schemes for systems described by a class of recurrent neural networks. We propose a procedure based on linear matrix inequalities for designing an observer and a static state-feedback controller. The algorithm leverages global and regional incremental input-to-state stability (incremental ISS) and enables the tracking of constant setpoints, ensuring robustness to disturbances and state estimation uncertainty. To address the potential limitations of regional incremental ISS, we introduce an alternative scheme in which the static law is replaced with a tube-based nonlinear model predictive controller (NMPC) that exploits regional incremental ISS properties. We show that these conditions enable the formulation of a robust NMPC law with guarantees of convergence and recursive feasibility, leading to an enlarged region of attraction. Theoretical results are validated through numerical simulations on the pH-neutralisation process benchmark.

eess.SY↗

Development of a velocity form for a class of RNNs, with application to offset-free nonlinear MPC design

This paper addresses the offset-free tracking problem for nonlinear systems described by a class of recurrent neural networks (RNNs). To compensate for constant disturbances and guarantee offset-free tracking in the presence of model-plant mismatches, we propose a novel reformulation of the RNN model in velocity form. Conditions based on linear matrix inequalities are then derived for the design of a nonlinear state observer and a nonlinear state-feedback controller, ensuring global or regional closed-loop stability of the origin of the velocity form dynamics. Moreover, to handle input and output constraints, a theoretically sound offset-free nonlinear model predictive control algorithm is developed. The algorithm exploits the velocity form model as the prediction model and the static controller as an auxiliary law for the definition of the terminal ingredients. Simulations on a pH-neutralisation process benchmark demonstrate the effectiveness of the proposed approach.

eess.SY↗