This package provides a library to create fuzzy inference systems in Ada.

For more information on fuzzy inference, see

https://en.wikipedia.org/wiki/Fuzzy_control_system and http://fr.mathworks.com/help/fuzzy/fuzzy-inference-process.html

To compile this library, you need to install the GNAT compiler and the GNAT Components Library (gnatcoll) first.

Then type::

make

See examples/example.adb for a simple demo that shows how to create a fuzzy inference engine programmatically. This demo simulates a thermostat that tries to smoothly maintain the temperature around 20 C, while at period interval we randomly nudge it.

Basically, we give the system rules like:

if temperature is COLD then power is FULL if temperature is MEDIUM then power is MEDIUM if temperature is HOT then power is LOW

then provide a temperature, and let the system compute what the output should be. We need a bit more setup, though, to define what the terms COLD, MEDIUM, HOT, FULL and LOW mean.

Type::

make
./examples/obj/example

to compile and run this demo

Another demo is provided. The goal is to control a robot so that it reaches its target while avoiding obstacles. The rules are described in an external file examples/robot.fcl, which you can modify before restarting the demo.

This file is written in the Fuzzy Controller Language, for which you can find a description at http://www.fuzzytech.com/binaries/ieccd1.pdf

This demo requires GtkAda on your system.

Type::

 make all robot

to compile and run this demo.