Robustness analysis of a paradigm for non-linear robot controllers based on soft computing techniques

Exact mathematical modelling and identification-based control of coupled non-linear MIMO systems still remained complicated and clumsy. VLSI technology also makes it possible to realize robust control based on "soft computing techniques" requiring only rough models but resulting in limited control quality with noisy fluctuations. The motion of quite simple mechanical systems as rigid bodies also may show chaotic behavior. Numerical techniques as finite-element methods in realization may act as an additional source of chaotic fluctuations. In this paper the relation between approximate modelling, robustness and chaotic nature of the controlled system is investigated via simulation. On this basis certain general rules are concluded regarding the operation of such controllers.<
>