Neural Network Based Disturbance Canceler with Feedback Error Learning for Nonholonomic Mobile Robots

  • Izumi, Kiyotaka (Department of Advanced Systems Control Engineering, Graduate School of Science and Engineering, Saga University) ;
  • Syam, Rafiuddin (Department of Advanced Systems Control Engineering, Graduate School of Science and Engineering, Saga University) ;
  • Watanabe, Keigo (Department of Advanced Systems Control Engineering, Graduate School of Science and Engineering, Saga University) ;
  • Kiguchi, Kazuo (Department of Advanced Systems Control Engineering, Graduate School of Science and Engineering, Saga University)
  • 발행 : 2003.09.01

초록

Conventional disturbance rejection methods have to derive the inverse model of a system. However, the inverse model of n nonholonomic system is not unique, because an inverse it changes depending on initial conditions and desired values. A kind of internal model control (IMC) using feedback error learning is discussed for the motion control of nonholonomic mobile robots in this paper, The present method is different from a conventional IMC whose control system consists of an inverse model, a direct model and a filter. The present disturbance rejection method need not use a direct model, where the remaining two elements are composed of the same inverse model based on neural networks.

키워드