Improvement of aeroelastic vehicles performance through recurrent neural network controllers

Brillante, Claudio; Mannarino, Andrea

doi:10.1007/s11071-015-2583-2

Cited by 9 publications

(8 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For multi-class change tasks, we use one-hot coding [30] to represent the labels Y of different change types. In the recent literature on the RNN framework [31], recent advances in statistical machine learning and deep learning [32] have yielded state-of-the-art results with powerful sequence models. In these models, an input sequence in an "end-to-end" fashion is given for both training and inference, and the models can maximize the probability of correctly-predicted results directly.…”

Section: Our Proposed Referee Modelmentioning

confidence: 99%

Learning a Transferable Change Rule from a Recurrent Neural Network for Land Cover Change Detection

2016

View full text Add to dashboard Cite

When exploited in remote sensing analysis, a reliable change rule with transfer ability can detect changes accurately and be applied widely. However, in practice, the complexity of land cover changes makes it difficult to use only one change rule or change feature learned from a given multi-temporal dataset to detect any other new target images without applying other learning processes. In this study, we consider the design of an efficient change rule having transferability to detect both binary and multi-class changes. The proposed method relies on an improved Long Short-Term Memory (LSTM) model to acquire and record the change information of long-term sequence remote sensing data. In particular, a core memory cell is utilized to learn the change rule from the information concerning binary changes or multi-class changes. Three gates are utilized to control the input, output and update of the LSTM model for optimization. In addition, the learned rule can be applied to detect changes and transfer the change rule from one learned image to another new target multi-temporal image. In this study, binary experiments, transfer experiments and multi-class change experiments are exploited to demonstrate the superiority of our method. Three contributions of this work can be summarized as follows: (1) the proposed method can learn an effective change rule to provide reliable change information for multi-temporal images; (2) the learned change rule has good transferability for detecting changes in new target images without any extra learning process, and the new target images should have a multi-spectral distribution similar to that of the training images; and (3) to the authors' best knowledge, this is the first time that deep learning in recurrent neural networks is exploited for change detection. In addition, under the framework of the proposed method, changes can be detected under both binary detection and multi-class change detection.

show abstract

Section: Our Proposed Referee Modelmentioning

confidence: 99%

Learning a Transferable Change Rule from a Recurrent Neural Network for Land Cover Change Detection

2016

View full text Add to dashboard Cite

show abstract

“…In the work of Wang et al (2011), a different NN was used to deal with scenarios under various exogenous disturbances. Brillante and Mannarino (2016) used two recurrent NNs for system identification and control. Tang et al (2018) made further improvement to AFS by deriving an algorithm powered by two NNs to synthesize nonlinear optimal control in real time.…”

Section: Introductionmentioning

confidence: 99%

A neural network approach for improving airfoil active flutter suppression under control-input constraints

Tang

Chen

Tian

et al. 2020

Journal of Vibration and Control

View full text Add to dashboard Cite

This study deals with improving airfoil active flutter suppression under control-input constraints from the optimal control perspective by proposing a novel optimal neural-network control. The proposed approach uses a modified value function approximation dynamically tuned by an extended Kalman filter to solve the Hamilton–Jacobi–Bellman equality online for continuously improved optimal control to address optimality in parameter-varying nonlinear systems. Control-input constraints are integrated into the controller synthesis by introducing a generalized nonquadratic cost function for control inputs. The feasibility of using a performance index involving the nonquadratic control-input cost with the modified value function approximation is examined through the Lyapunov stability analysis. Wind tunnel experiments were conducted for controller validation, where an optimal controller synthesized offline via linear parameter-varying technique was used as a benchmark and compared. It is shown, both theoretically and experimentally, that the proposed method can effectively improve airfoil active flutter suppression under control-input constraints.

show abstract

“…Also, the neural network-based approach were proposed for aeroelastic problems, e.g. (Voitcu and Wong, 2003; Hongkun et al , 2017), although mainly in the context of flutter suppression, see Bernelli-Zazzera et al (2000); Brillante and Mannarino (2016), Mattaboni et al (2009).…”

Section: Introductionmentioning

confidence: 99%

Suppression of nonlinear aeroelastic vibrations by learned neural network controller

Dul

2019

AEAT

View full text Add to dashboard Cite

Purpose The purpose of the paper is to analyze the active suppression of the aeroelastic vibrations of ailerons with strongly nonlinear characteristics by neural network/reinforcement learning (NN/RL) control method and comparing it with the classic robust methods of suppression. Design/methodology/approach The flexible wing and aileron with hysteresis nonlinearity is treated as a plant-controller system and NN/RL and robust controller are used to suppress the nonlinear aeroelastic vibrations of aileron. The simulation approach is used for analyzing the efficiency of both types of methods in suppressing of such vibrations. Findings The analysis shows that the NN/RL controller is able to suppress the nonlinear vibrations of aileron much better than linear robust method, although its efficiency depends essentially on the NN topology as well as on the RL strategy. Research limitations/implications Only numerical analysis was carried out; thus, the proposed solution is of theoretical value, and its application to the real suppression of aeroelastic vibrations requires further research. Practical implications The work shows the NN/RL method has a great potential in improving suppression of highly nonlinear aeroelastic vibrations, opposed to the classical robust methods that probably reach their limits in this area. Originality/value The work raises the questions of controllability of the highly nonlinear aeroelastic systems by means of classical robust and NN/RL methods of control.

show abstract

Improvement of aeroelastic vehicles performance through recurrent neural network controllers

Cited by 9 publications

References 49 publications

Learning a Transferable Change Rule from a Recurrent Neural Network for Land Cover Change Detection

Learning a Transferable Change Rule from a Recurrent Neural Network for Land Cover Change Detection

A neural network approach for improving airfoil active flutter suppression under control-input constraints

Suppression of nonlinear aeroelastic vibrations by learned neural network controller

Contact Info

Product

Resources

About