Complete controllability of continuous-time recurrent neural networks

Eduardo Sontag, Héctor Sussmann

Research output: Contribution to journalArticlepeer-review

35 Scopus citations

Abstract

This paper studies controllability for the class of control systems commonly called (continuous-time) recurrent neural networks. It is shown that, under a generic condition on the input matrix, the system is controllable, for every possible state matrix. The result holds when the activation function is the hyperbolic tangent.

Original languageEnglish (US)
Pages (from-to)177-183
Number of pages7
JournalSystems and Control Letters
Volume30
Issue number4
DOIs
StatePublished - May 1997

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • General Computer Science
  • Mechanical Engineering
  • Electrical and Electronic Engineering

Keywords

  • Global stabilization
  • Linear discrete-time systems
  • Saturated feedback

Fingerprint

Dive into the research topics of 'Complete controllability of continuous-time recurrent neural networks'. Together they form a unique fingerprint.

Cite this