We consider a problem, that is basic to perception measurement: an artifici
al system that mimics the integration of perceptual information into a prop
osed framework performed by the human senses. The paper presents a method f
or optimising complex information from one single sensor that is indirectly
based on the human measurement capability, in the mixing process of dough.
We propose a method to measure, analyse and correlate sensor information.
This technique is exemplified by an industrial application, where the influ
ence of the properties of the human senses on the rheological behaviour of
wheat flour dough is estimated and has further been studied in the dynamic
mixing process. It is shown that, by dynamic measurements on the mixing mac
hine, the complex behaviour of the mixing process could clearly be followed
. However, by implementing human knowledge and experience into the sensor s
ystem we manage to adapt a process that works with human preferences. The s
ystem's ability to optimise various types of bread formulae and different t
ypes of wheat flour is also described. An additional benefit is the possibi
lity, by calculations, to predict the behaviour of the wheat flour dough pr
ocess. (C) 1999 Elsevier Science Ltd. All rights reserved.