Recurrent Neural Network Controller for Linear and Nonlinear Systems

Sheikhmemari S.

4th International Conference on Intelligent and Fuzzy Systems (INFUS), Bornova, Turkey, 19 - 21 July 2022, vol.504, pp.752-760 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 504
  • Doi Number: 10.1007/978-3-031-09173-5_86
  • City: Bornova
  • Country: Turkey
  • Page Numbers: pp.752-760
  • Keywords: Recurrent neural network self-tuning PID controller, Deep learning recurrent neural network
  • Yıldız Technical University Affiliated: No

Abstract

This paper aims to design deep recurrent neural network controllers for linear and non-linear systems. Design and comparison with two different controllers are considered. A pure recurrent neural network controller and self-tuning PID controller based on recurrent neural networks (PRNN), according to the influence of the object's parameter on system output performance, The PRNN can auto-adjust its weights to vary k(P), k(I) and k(D) The emulation results show that the presented control systems have quick response speed and strong adaptive capability.