Fuzzy component network for intelligent measurement and control

摘要

This paper presents the concept of fuzzy component networks. First, we introduce fuzzy components such as fuzzy sensors, fuzzy actuators, and fuzzy inference components. Then, we describe the implementation of such components into a generic component, called a fuzzy cell. Depending on its configuration, a fuzzy cell is able to perform the functions of one or several fuzzy components. Each fuzzy cell is implemented by means of the Intel 80C196 microcontroller and is connected to a field bus. Fuzzy cells are characterized by their ability to handle and exchange fuzzy symbolic information. Their configuration is performed via the field bus using the language prototype of integrated language for symbolic sensors and actuators (PLICAS) that was specially designed for this purpose. A PLICAS compiler is integrated inside each cell, thus conferring interoperability properties to the cell. Two applications are presented: 1) a fuzzy controller whose fuzzification, inference, end defuzzification interfaces are distributed into three fuzzy cells connected by the fieldbus; and 2) the fusion of information provided by two ultrasonic fuzzy sensors for the detection of obstacles.