A data-based neural controller training method with tunable stability margin using multi-objective optimization

This paper presents a method to design neural network controllers based on imitation learning and with tunable stability guarantees through multi-objective optimization. Stability margins are derived from analyzing state-space neural networks based on the representation of nonlinear activation funct...

Full description

Saved in:
Bibliographic Details
Published in:IFAC-PapersOnLine 2023-01, Vol.56 (2), p.3092-3099
Main Authors: Pinguet, Jérémy, Feyel, Philippe, Sandou, Guillaume
Format: Article
Language:English
Subjects:
Imitation learning
Multi-objective optimization
State-space neural networks
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper presents a method to design neural network controllers based on imitation learning and with tunable stability guarantees through multi-objective optimization. Stability margins are derived from analyzing state-space neural networks based on the representation of nonlinear activation functions by linear parameter varying models. The controller training is formulated as a multi-objective problem whose solutions yield a set of the best trade-offs in terms of minimal imitation error and maximal stability margins. The proposed approach is illustrated in the synthesis by imitation of an aircraft neural autopilot using a flight simulator.
ISSN:2405-8963
2405-8963
DOI:10.1016/j.ifacol.2023.10.1440