Нашли опечатку? Выделите ее мышкой и нажмите Ctrl+Enter
Название: Neural network control of nonlinear discrete-time systems
Автор: Sarangapani J.
Modern feedback control systems have been responsible for major successes
in the fields of aerospace engineering, automotive technology, defense, and
industrial systems. The function of a feedback controller is to alter the behavior
of the system in order to meet a desired level of performance. Modern control
techniques, whether linear or nonlinear, were developed using state space or frequency
domain theories. These techniques were responsible for effective flight
control systems, engine and emission controllers, space shuttle controllers,
and for industrial systems. The complexity of today’s man-made systems has
placed severe constraints on existing feedback design techniques. More stringent
performance requirements in both speed and accuracy in the face of system
uncertainties and unknown environments have challenged the limits of modern
control. Operating a complex system in different regimes requires that the controller
be intelligent with adaptive and learning capabilities in the presence of
unknown disturbances, unmodeled dynamics, and unstructured uncertainties.
Moreover, these controllers driven by the hydraulic, electrical, pneumatic, and
bio-electrical actuators have severe multiple nonlinearities in terms of friction,
deadzone, backlash, and time delays.