BISC Seminar
Joseph AGUILAR-MARTIN
Directeur
Laboratoire d'Automatique et d'Analyse des Systems
7 Av Colonel Roche
31077 Touluse Cedex, France
33 61 33 64 74 tel
33 61 55 35 77 fax
19 September 1996
310 Soda Hall
4:00-5:00pm
Abstract:
In the real world, any real regulation problems depends on the energy
disponibility, introducing natural saturations in the possibilities of the
control actions. The "Natural" apûproach consists in the normalization of
the control action signal in the unit interval so that the saturation
constraints are allways respected. By doing so the control signal becomes a
mapping between the vector output signal and the unit interval. If for each
component of the output signal a "marginal" control can be defined, the
controller must perform a combination of marginal control signals, all of
them being defined in the unit interval. NLC (Natural Logic Control
approach) takes advantage of that situation to imbed regulation design into
fuzzy logic compound implication.
SIMO (single input, multi-output) systems are difficult to be handled by
simple fuzzy controllers, because of the dimensionnality of the non-linear
hypersurface developped by such regulators The NLC overcomes this difficulty
by considering the connexion of as many outputs as necessary as a fuzzy
logical combination using t-norms and t-conorms.
The two extreme attitudes being the conjunction (AND) that leads to the
sharper result and the disjunction(OR) that leads to the milder one, it
appears convenient to compensate both attitudes by means of a linear
interpolation depending on a sinle parameter l in the unit interval.
In previous works it has been proved that if a De Morgan pair of fuzzy
connectives is choses, the properties of the result is completely ordered
with respect to l. This enables in practice a simple adaptivie procedure
that adjustes this parameter according to the desired performances of the
closed-loop system.
Simulation results in Matlab can be shown and commented.
-- _______________________________________________________________________ Michael A. Lee Post Doctoral Researcher Berkeley Initiative in Soft Computing 387 Soda Hall Tel: +1-510-642-9827 Computer Science Division Fax: +1-510-642-5775 University of California Email: leem@cs.berkeley.edu Berkeley, CA 94720-1776 USA WWW: http://www.cs.berkeley.edu/~leem _______________________________________________________________________