Fuzzy Linguistic Methods for the Aggregation of Complementary Sensor Information

Abstract

The problem of the aggregation of complementary information is a crucial point in the monitoring of large intelligent systems. This paper deals with the cases in which there is no analytical mathematic model to derive new information from the basic measurements. Our artificial intelligence approach consists in using linguistic knowledge provided by experts. In the second section, we explain within the framework of the fuzzy subset theory how the numeric and linguistic representations could be used in the measurement aggregation problems. Next, the third section proposes an interpolation mechanism that creates a fuzzy partition of the numeric multi-dimensional space of the basic features. In section four, we present a formal symbolical representation of fuzzy If ... Then ... rules for the combination of basic features. The proposed methods are then applied in section five to the problem of the navigation of a mobile robot equipped with two ultrasonic range finding sensors, giving the proximity, the orientation, and the danger of an obstacle by the aggregation of the distance information provided by each one.