2-Input-2-Output Fuzzy Logic Based Hemodynamic Controller
Applied to a Non-Linear Canine Circulatory System Model

Non-Linear Circulatory Model developed by Clement Yu
Fuzzy Logic Controller by Claudio M. Held
Pascal-to-C Translation by Johnnie W. Huang

Last Revised on Feb. 8, 1995

Please address all correspondences to:

Dr. Howard Kaufman
Department of Electrical, Computer, and System Engineering
Rensselaer Polytechnic Institute
Troy, NY 12180 USA
kaufmh@rpi.edu

Dr. Rob J. Roy
Department of Biomedical Engineering
Rensselaer Polytechnic Institute
Troy, NY 12180 USA
royr@rpi.edu

This text file contains information on customizing the enclosed program
for testing other prototype hemodynamic controllers.

Please refer to the attached README.TXT file for the specifications
and descriptions on compiling and executing this program.

Remarks:
- Line numbers indicated by a set of parenthesis refers to the location
of a particular function or a variable within the program, i.e. (L#235)
refers to line 235 in the program where it may be modified according
to the user's customized specification.

----------------

Following proper compilation and execution of the codes, the system
will first initiate (L#472-516) itself with the default parameters.
The user has the option to re-define the initial states of the subject
(e.g. cardiac contractility) by loading a replacement set of circulatory
parameters (L#2793).  These are the files which have the extensions
*.nlm (L#518)

As the simulation proceeds, the pressure and the blood flow in each
compartments will be calculated using the Euler algorithm having 180
integration steps (L#150) per heart beat which will compute the
new volume in each compartments (L#2939).

The embedded fuzzy logic controller intervenes (L#701) every 30 seconds
(L#152) to monitor the mean arterial pressure (MAP) and cardiac output
(CO) of the subject.  It currently has two modes: supervisory and fuzzy
logic mode.  The supervisory mode (L#1890) is activated when either MAP
or CO falls outside of the fuzzy ranges, typically when the patient
is in critical situation.  The fuzzy logic mode (L#2060) is used to
lead the hemodynamic variables to the target zone pre-defined in
the control parameter set (L#1802).  Fuzzification of the values are carried
out in FUZZIF (L#2497).  A fuzzy rule table (L#832-1388) is used to
evaluate for fuzzified drug dosages based on the current fuzzified
hemodynamic states.  Defuzzification process in done in DEFUZZ (L#2561).

A new controller may be substituted in place of FUZ_SUPERV_CONTROL (L#1848).
The output of the controller should be the infusion rates of sodium
nitroprusside (SNP) and dopamine (DPM).  Other drugs may also be
added or substituted.

The concentration of the drugs are updated in DRUG_UPDATE (L#3254) called
from the simulation process (L#650) and the function  UPDATE_ALL_CONC
(L#2845) will be used to determine the corresponding concentration
in each compartments.  The effects of the drugs are incorporated in the
model in the function UPDATE_BEAT_PARAMS (#L2864).  Subject sensitivity
to a given drug may be modified by tagging an extra parameter to the
computation of drug effects (#L499-510) during the iteration of the
concentration-effect differential equation. (#L2888-2892)



----------------

REFERENCES on the Non-Linear Model and/or the Fuzzy Logic Controller:

Gopinath R, Bequette BW, Kaufman H, Held CM, Huang JW, Roy RJ.  "Drug Delivery
Control System Challenge", Proc. of IEEE Conf. on Decision and Control, 1995.

Held CM, Roy RJ, "Multiple Drug Hemodynamic Control by Means of a S
upervisory-Fuzzy Rule Based Adaptive Control System:  Validation on a
Model," IEEE Trans. Biomed. Eng., April 1995, pp371-385

Yu C, Roy RJ, Kaufman H. "A Circulatory Model for Combined Nitroprusside-
Dopamine Therapy", Medical Progress Through Technology, 16:77-88, Kluwer
Academic Publishing,1990