| Abstract |
Speed and position control of electric drive systems are often made difficult by
torsional vibrations, nonlinearities and partially unknown system structure and parameters.
A possible solution approach is stable model reference neurocontr'ol, a certain class
of neural network and neuro-fuzzy control concepts which have the particular advantage
of mathematically guaranteed stability and performance. In this paper some recent developments
of stable model reference neurocontrol are summarized, and its applicability
for controlling electric drive systems with torsional vibrations is evaluated by simulation
and experimental examples. |
