Advanced control based on Recurrent Neural Networks learned using Virtual Reference Feedback Tuning and application to an Electronic Throttle Body (with supplementary material)

TL;DR

This paper explores the use of Virtual Reference Feedback Tuning to train recurrent neural network controllers, specifically ESN and LSTM, for nonlinear systems like electronic throttle bodies, achieving zero steady-state error.

Contribution

It introduces a novel control scheme using VRFT with RNNs to control nonlinear systems and demonstrates its effectiveness on an electronic throttle body benchmark.

Findings

RNN controllers can effectively constrain control variables.

The proposed method achieves zero steady-state error.

Validation on an electronic throttle body confirms practical applicability.

Abstract

In this paper the application of Virtual Reference Feedback Tuning (VRFT) for control of nonlinear systems with regulators defined by Echo State Networks (ESN) and Long Short Term Memory (LSTM) networks is investigated. The capability of this class of regulators of constraining the control variable is pointed out and an advanced control scheme that allows to achieve zero steady-state error is presented. The developed algorithms are validated on a benchmark example that consists of an electronic throttle body (ETB).