Fuzzy internal models in vision systems modelling

Abstract

The authors study internal-model representations in the framework of control theory and of visual system modelling. It is shown that the control of rapid orienting movements of the visual axis in space (saccadic system) can be modelled as a control system based on internal models. Moreover, a fuzzy approach is found to be very well adapted for the modelling of some parts of the saccadic system (related to highly nonlinear sensors and/or internal representations). Classical properties of this control structure such as robustness and disturbance rejection are met.<>