HIGH-SPEED ACTIVE GAS ODOR SENSING SYSTEM USING ADAPTIVE-CONTROL THEORY/

An active gas/odor sensing system using an internal blender was previo usly proposed to measure the mixture composition of an aroma. In the s ystem, the mixture composition of the blender outlet was repeatedly ad justed using an optimization algorithm to match the sensor array outpu t pattern of blended vapor to that of a test vapor. The system is high ly flexible and can be used even when sensors have nonlinearities and the additive property is not valid. Since vapor and air were repeatedl y supplied to the sensor array, the previous measurement was lengthy. Here, the concentration of each component vapor in the blender is cont inuously changed using adaptive feedback control theory so that the me asurement time can be reduced. After the introduction of adaptive cont rol theory, the measurement was drastically speeded up. Using this app roach, the quantification results of both two-component and three-comp onent vapors were successfully obtained within at least 2 min, whereas more than 40 min was required in the previous method. (C) 1997 Elsevi er Science S.A.