LEARNING AND OUTPUT REGULATION WITH RECURRENT NEURONAL NETWORKS
Jorge M. O. Henriques, B. Castillo, André Titli, Paulo Sousa Gil, Antonio Carlos Dourado · 2000
The problem of on-line learning using recurrent neural networks, using within a model based non- linear control scheme, is addressed in this paper. The control output regulation theory is introduced, leading to an indirect adaptive control structure and matching both the classical non-linear control and neural methods. The main goal is to investigate the combination of the well-known learning abilities of neural networks with the stability properties of the output regulator. The practical potentials of the proposed scheme are demonstrated by some experimental results on a laboratory bench process.