23 Hits in 0.77 sec

On the adaptation of recurrent neural networks for system identification [article]

Marco Forgione, Aneri Muni, Dario Piga, Marco Gallieri
2022 arXiv   pre-print
Acknowledgments The activities of Marco Forgione and Dario Piga have been supported by HASLER STIFTUNG under the project DEALING: DEep learning for dynamicAL systems and dynamical systems for deep learnING  ... 
arXiv:2201.08660v1 fatcat:nfx44ocatjfavcjjqk2ulzqany

Infinite-Horizon Differentiable Model Predictive Control [article]

Sebastian East, Marco Gallieri, Jonathan Masci, Jan Koutnik, Mark Cannon
2020 arXiv   pre-print
Model Predictive Control (MPC, Maciejowski, 2000; Camacho & Bordons, 2007; Rawlings & Mayne, 2009; Kouvaritakis & Cannon, 2015; Gallieri, 2016; Borrelli et al., 2017; Raković & Levine, 2019) is the most  ... 
arXiv:2001.02244v1 fatcat:yns5l7dnwzg47al36tuewewj5m

Model predictive control with prioritised actuators

Marco Gallieri, Jan M. Maciejowski
2015 2015 European Control Conference (ECC)  
Research supported by the EPSRC grant "Control for Energy and Sustainability", EP/G066477/1. 1 Marco Gallieri is with McLaren Racing Limited, Chertsey Road, Woking GU21 4YH, UK Jan  ... 
doi:10.1109/ecc.2015.7330598 dblp:conf/eucc/GallieriM15 fatcat:737cylztxna25ig2dykfqixniu

Tustin neural networks: a class of recurrent nets for adaptive MPC of mechanical systems [article]

Simone Pozzoli, Marco Gallieri, Riccardo Scattolini
2019 arXiv   pre-print
The use of recurrent neural networks to represent the dynamics of unstable systems is difficult due to the need to properly initialize their internal states, which in most of the cases do not have any physical meaning, consequent to the non-smoothness of the optimization problem. For this reason, in this paper focus is placed on mechanical systems characterized by a number of degrees of freedom, each one represented by two states, namely position and velocity. For these systems, a new recurrent
more » ... neural network is proposed: Tustin-Net. Inspired by second-order dynamics, the network hidden states can be straightforwardly estimated, as their differential relationships with the measured states are hardcoded in the forward pass. The proposed structure is used to model a double inverted pendulum and for model-based Reinforcement Learning, where an adaptive Model Predictive Controller scheme using the Unscented Kalman Filter is proposed to deal with parameter changes in the system.
arXiv:1911.01310v1 fatcat:moyed5aloregzmjibzzzto5hay

SNODE: Spectral Discretization of Neural ODEs for System Identification [article]

Alessio Quaglino, Marco Gallieri, Jonathan Masci, Jan Koutník
2020 arXiv   pre-print
This paper proposes the use of spectral element methods for fast and accurate training of Neural Ordinary Differential Equations (ODE-Nets; ) for system identification. This is achieved by expressing their dynamics as a truncated series of Legendre polynomials. The series coefficients, as well as the network weights, are computed by minimizing the weighted sum of the loss function and the violation of the ODE-Net dynamics. The problem is solved by coordinate descent that alternately minimizes,
more » ... ith respect to the coefficients and the weights, two unconstrained sub-problems using standard backpropagation and gradient methods. The resulting optimization scheme is fully time-parallel and results in a low memory footprint. Experimental comparison to standard methods, such as backpropagation through explicit solvers and the adjoint technique , on training surrogate models of small and medium-scale dynamical systems shows that it is at least one order of magnitude faster at reaching a comparable value of the loss function. The corresponding testing MSE is one order of magnitude smaller as well, suggesting generalization capabilities increase.
arXiv:1906.07038v2 fatcat:5hspthjo5jgbdacluxtfbsiaga

Altitude control feasibility for a seaweed harvester

Marco Gallieri, John Ringwood
2010 2010 IEEE International Conference on Industrial Technology  
The output matrix, (x) ∈ ℝ 14×84 , in (10), is given by: (x) = ⎡ ⎣ 0 0 0 0 0 0 (x) ⎤ ⎦ (30) where and (x), are given in Gallieri [13] .  ...  The complete system dynamics Expanding the formulation of Fossen and Perez [5] , a 2body unconstrained 12 DOF motion model can be obtained, as in Gallieri [13] .  ... 
doi:10.1109/icit.2010.5472654 fatcat:oeyt6hjm6zflzeesnqsdsbj7ce

Real-time Classification from Short Event-Camera Streams using Input-filtering Neural ODEs [article]

Giorgio Giannone, Asha Anoosheh, Alessio Quaglino, Pierluca D'Oro, Marco Gallieri, Jonathan Masci
2020 arXiv   pre-print
Event-based cameras are novel, efficient sensors inspired by the human vision system, generating an asynchronous, pixel-wise stream of data. Learning from such data is generally performed through heavy preprocessing and event integration into images. This requires buffering of possibly long sequences and can limit the response time of the inference system. In this work, we instead propose to directly use events from a DVS camera, a stream of intensity changes and their spatial coordinates. This
more » ... sequence is used as the input for a novel asynchronous RNN-like architecture, the Input-filtering Neural ODEs (INODE). This is inspired by the dynamical systems and filtering literature. INODE is an extension of Neural ODEs (NODE) that allows for input signals to be continuously fed to the network, like in filtering. The approach naturally handles batches of time series with irregular time-stamps by implementing a batch forward Euler solver. INODE is trained like a standard RNN, it learns to discriminate short event sequences and to perform event-by-event online inference. We demonstrate our approach on a series of classification tasks, comparing against a set of LSTM baselines. We show that, independently of the camera resolution, INODE can outperform the baselines by a large margin on the ASL task and it's on par with a much larger LSTM for the NCALTECH task. Finally, we show that INODE is accurate even when provided with very few events.
arXiv:2004.03156v1 fatcat:spus7jl5wbhjxkfwio3ce723eq

Terminal spacecraft rendezvous and capture with LASSO model predictive control

Edward N. Hartley, Marco Gallieri, Jan M. Maciejowski
2013 International Journal of Control  
This paper explores how the recently investigated l 1 -regularised MPC (Ohlsson et al. 2010 , Nagahara and Quevedo 2011 , Annergren et al. 2012 , Gallieri and Maciejowski 2012 , 2013 ) -which amalgamates  ...  The recently considered l 1 regularised ( asso ) cost function for MPC (Ohlsson et al. 2010 , Nagahara and Quevedo 2011 , Annergren et al. 2012 , Gallieri and Maciejowski 2012 , 2013 could also be an  ... 
doi:10.1080/00207179.2013.789608 fatcat:l6noeaslmzgwlg2upbjqk5pbxa

Safe Interactive Model-Based Learning [article]

Marco Gallieri and Seyed Sina Mirrazavi Salehian and Nihat Engin Toklu and Alessio Quaglino and Jonathan Masci and Jan Koutník and Faustino Gomez
2019 arXiv   pre-print
Vinogradska ; Rawlings & Mayne (2009); Kouvaritakis & Cannon (2015); Gallieri (2016); Borrelli et al. (2017); Raković & Levine (2019).  ... 
arXiv:1911.06556v2 fatcat:fx22zrnlhfdsxfqen66utzk3zi

Predictor Design for Altitude Control of a Seaweed Harvester

Marco Gallieri, John Ringwood, Andrea Giantomassi, Gianluca Ippoliti, Sauro Longhi
2010 IFAC Proceedings Volumes  
The control performances are then evaluated, and the results are compared to Gallieri and Ringwood (2010) .  ...  The control approach of Gallieri and Ringwood (2010), including a feedforward action, which requires a single step disturbance prediction, is investigated further, focusing on the disturbance prediction  ...  and Ringwood (2010) ; Gallieri (2009a,b) and reference therein.  ... 
doi:10.3182/20100915-3-de-3008.00006 fatcat:jmaqgd5fffg6fb3qf3jrtf6n5m

Triangular formation control using range measurements: An application to marine robotic vehicles

Jorge M. Soares, A. Pedro Aguiar, António M. Pascoal, Marco Gallieri
2012 IFAC Proceedings Volumes  
This paper addresses the problem of maintaining an autonomous robotic vehicle in a moving triangular formation by regulating its position with respect to two leader vehicles. The robotic vehicle has no a priori knowledge of the path described by the leaders and its goal is to follow them by constantly regulating the inter-vehicle distances to a desired fixed value, using range-only measurements. To solve this station keeping problem, we propose a control strategy that estimates the formation
more » ... ed and heading from the ranges obtained to the two leading vehicles, and uses simple feedback laws for speed and heading commands to drive suitably defined common and differential errors to zero. For straight-line motion, we provide guaranteed conditions under which the proposed control strategy achieves local convergence of the distance errors to zero. We also indicate how our design procedure can be extended to full dynamic models of marine robotic vehicles equipped with inner loops for yaw and speed control. Simulation results using realistic models are described and discussed.
doi:10.3182/20120410-3-pt-4028.00020 fatcat:pu7q5eu52na43axa2kjvhvidcm

Neural Lyapunov Model Predictive Control: Learning Safe Global Controllers from Sub-optimal Examples [article]

Mayank Mittal, Marco Gallieri, Alessio Quaglino, Seyed Sina Mirrazavi Salehian, Jan Koutník
2021 arXiv   pre-print
Correspondence to: Mayank Mittal <>, Marco Gallieri <>. using Lyapunov functions.  ...  Neural Lyapunov Model Predictive Control Supplementary Material Mayank Mittal * 1 2 Marco Gallieri * 1 Alessio Quaglino 1 Seyed Sina Mirrazavi Salehian 1 Jan Koutník 1 A.  ... 
arXiv:2002.10451v2 fatcat:mpmbdjk6ozhsbkqptyl2zl55oq

NAIS-Net: Stable Deep Networks from Non-Autonomous Differential Equations [article]

Marco Ciccone, Marco Gallieri, Jonathan Masci, Christian Osendorfer, Faustino Gomez
2021 arXiv   pre-print
This paper introduces Non-Autonomous Input-Output Stable Network(NAIS-Net), a very deep architecture where each stacked processing block is derived from a time-invariant non-autonomous dynamical system. Non-autonomy is implemented by skip connections from the block input to each of the unrolled processing stages and allows stability to be enforced so that blocks can be unrolled adaptively to a pattern-dependent processing depth. NAIS-Net induces non-trivial, Lipschitz input-output maps, even
more » ... an infinite unroll length. We prove that the network is globally asymptotically stable so that for every initial condition there is exactly one input-dependent equilibrium assuming tanh units, and incrementally stable for ReL units. An efficient implementation that enforces the stability under derived conditions for both fully-connected and convolutional layers is also presented. Experimental results show how NAIS-Net exhibits stability in practice, yielding a significant reduction in generalization gap compared to ResNets.
arXiv:1804.07209v4 fatcat:4ybqqnnlu5cjbp73mxq3yuacba

Page 22 of Musical World Vol. 33, Issue 2 [page]

1855 Musical World  
22 THE MUSICAL WORLD. still he is a good singer; as an actor, he is wanting in animation’ The baritone, Signor Ferri (Marco Visconti), is an artist o decided merit.  ...  Gallieri, was given, interpreted by the Signora Cominotti, and the Signori Biondi and Ferrari. It would be unjust to criticise the new work until the execution is improved.  ... 

Ensino elementar no Brasil e na Itália: o caso dos imigrantes italianos na escola de Cascatinha (Petrópolis, Estado do Rio de Janeiro)

Carlo Pagani
2014 Acta Scientiarum : Education  
O Padre Dom Carlo Gallieri, honorário da Sociedade, juntamente com o Prof.  ...  De Pescatina, aldeia situada na região de Verona, vieram quase todos os moradores, inclusive o padre Dom Carlo Gallieri 4 , viajando no mesmo navio.  ... 
doi:10.4025/actascieduc.v36i2.22241 fatcat:u6zcnavz6japxevluv4y7irbwi
« Previous Showing results 1 — 15 out of 23 results