For the modelling, feedforward and feedback control of nonlinear processes one uses frequently two- and/or multidimensional characteristics. Such characteristics can be represented effectively by locally generalizing storage methods. By three different examples and their further developments it is shown in this paper, that very different approaches are possible and that taking into account the inherent advantages and disadvantages one should choose the most appropriate method according to the envisaged application. Part I deals through CMAC/AMS with a neurobiologically motivated technique and through MIAS with one derived from the area of computer graphics.
Print ISSN: 0178-2312
Volume: 52, 03/2004
Pages: 117