|
The Control Toolbox of 20-sim contains several tools that can
aid you in developing controllers for your modeled machines, the
Controller Design Editor, the Filter Editor and The Neural Network
Editors.
Controller Design Editor
The controller design editor is a specialized tool for the design
of feedback control systems. A feedback structure of subsystems
is presented with a linear plant, controller, measurement and
pre-filter. Also the open-loop and closed-loop gains and the sensitivities
are available.
The Controller Design Editor.
You can edit your controller as an ABCD State Space system, a
Transfer Function or in a Zero Pole Gain form. Changes in one
of the subsystems directly update all open plots and dialogs.
For instance, adapting the controller gain immediately changes
poles and zeros of the closed-loop system and the overall step
response. The integration within 20-sim and linear system exchange with
MATLAB makes this editor a powerful tool for designing feedback
control systems!
Filter Editor
With the Filter editor you can create your own linear filters
according to your specifications. Available filters are Bessel,
Butterworth and ChebyChev filters where you can specify the order
and characteristic frequencies. A choice can also be made from
PID, lead/lag, or notch filters.
The Filter Editor.
Neural Network Editors
The 20-sim Control toolbox supports two well-known networks:
Adaptive B-Spline Networks and Multi-Layer Perceptron Networks.
The B-Spline Editor.
These neural networks must be trained by repeatedly presenting
examples to the network. Each example includes both inputs and
desired outputs. Based on the error between desired outputs and
the real network output, the neural network adapts the weight
of each neuron according to a user-defined learning rate.
If the response is accurate enough, you can save the weights of
the neurons, to use the neural network in your controller. An
unlimited number of neurons may be used.
|
