Adaptive control

The work in this area, extending over several years, has been mainly based on the model-reference approach, and principally concerned with the robust properties of the algorithms used, in the presence of disturbances and unmodelled dynamics.

The use of neural networks for the adaptive control of non-linear systems has also been considered, particularly by training a network on-line as it attempts to undertake a control task. In addition, work has also been done on developing systems which can imitate the human ability to learn by practice how to perform a control task effectively. As an alternative to neural networks or fuzzy logic, this approach employs a technique known as vector field modelling, in which the system builds up a model of the unknown plant structure and dynamics from data acquired during operation.

  • P.A. Cook "Application of Model Reference Adaptive Control to a Benchmark Problem", Automatica, 30 (1994), 585-588.
  • G.W. Ng & P.A. Cook "Real-Time Control of Systems with Unknown and Varying Time-Delays using Neural Networks", Engineering Applications of Artificial Intelligence, 11 (1998), 401-409
▲ Up to the top