readme.txt

非线性控制 Matlab编译

Nonlinear control algorithm toolbox.

Files to implement nonlinear feed-back control of dynamical systems.
Full description of the algorithm can be found in
``Model-Independent Nonlinear Control Algorithm with Application 
to a Liquid Bridge Experiment'' by Valery Petrov, Anders Haaning, 
Kurt A. Muehlner, Stephen J. Van Hook, and Harry L. Swinney,
to appear in Phys. Rev. E (1998).   For brief description see
V.Petrov, K. Showalter, "Nonlinear Control from Time-Series" 
Phys Rev Lett 76, 3312 (1996).

contr.m is the front end for the algorithm. Please read the
usage by typing 'help contr' in Matlab.

The demonstration files in this toolbox are intended as guides to
using the control algorithm.  In each example, unstable steady states
are targeted by the control alogorithm through calls to contr.m. 
It is the hope of the authors that by running the example programs and 
reading through the corresponding code, anyone wishing to use the algorithm
will be able to do so with a minimum of difficulty.

Run logdemo.m for a demonstration of stabilizing the fixed point of
the logistic map:  x(i+1) = mu  * x(i) * [1 - x(i)],
where mu can run from 0 to 4.  Changing mu changes the system dynamics --
from a single fixed point to chaos (see one of the many books that 
discuss the logistic map for more discussion); mu is set in the
logsim.m program.  As currently implemented, the logistic map is in the
stable period four regime (mu = 3.5).  

Run lordemo.m for a demonstration of targeting and switching between unstable states
in the chaotic Lorenz system. 
% note: Sometime it stops right after the start. Type lordemo again to start it.

The file pendemo.m demonstrates using the algorithm to target a specific state,
in this case the unstable upright position of a simple pendulum.  
% note: pendemo4 optimized for Matlab 4.2, pendemo5 for Matlab 5

Run brdemo.m for a demonstration of using the algorithm to 
stabilize an unstable state in a MIMO (two-input-two output) 
4-dimensional nonlinear system, which models hydrodynamical
instabilities in a liquid bridge.  This example is meant to show
the use of the algorithm in typical laboratory conditions:  the presence of
noise in measurements, limitations on feedback capabilities, and the
use of time delayed coordinates to deal with insufficient system determination
are all modeled.
 
For additional information on the liquid bridge refer to:
http://chaos.ph.utexas.edu/~lera/lb.html

05-01-98 Val Petrov CNLD, University of Texas, Austin
Please, send all the questions to [e-mail]

Copyright (c) 1998 The University of Texas at Austin