The perceiving sensory estimated in an artificial human estimation based sensor system

Author(s): Winquist, F., Lauber, A.

Journal/Book: IMTC/97 IEEE INSTRUMENTATIO. 1997; 345 E 47TH ST, NEW YORK, NY 10017. Order Info: I E E E SERVICE CENTER, 445 HOES LANE, PO BOX 1331, PISCATAWAY, NJ 08855-1331 USA, 1509 pp., 294 chaps., Plates, PRICE NOT AVAILABLE (hardbound) $210.00 SET PRICE $105.00 MEMBER PRICE (softbound) Host site: OTTAWA, CANADA Sponsors: IEEE Instrumentat & Measurement Soc, IEEE Ottawa Sect. I E E E SERVICE CENTER. 570-575.

Abstract: We consider a problem which is basic to perception measurement: to devise an artificial system that mimics the integration of perceptual information into a proposed framework performed by the human sensory. This paper presents a method for handling complex perceptual information from different sensors that imitate human appreciation and enjoyment of food, in our case potato chips. We propose a method to measure, analyze and correlate information from a ''crush chamber''. Different sensors measure auditory, olfactory and tactile data, then analyze and merge the information into a system opinion. The proposed technique is exemplified by two applications, to estimate aging and to classify different potato chips by collecting, analyzing and fusing sensory quality parameters: crunching, hearing and smelling information. It is shown that by analyzing each single quality parameter, the aging process of the chips could not clearly by followed. However, fusing multivariate analyzis (PCA) of the total sample information we are able to follow the aging process. We are also able to classify various types of chips.

Note: Article Wide P, Linkoping Univ, Dept Phys & Measurement, S-58183 Linkoping, SWEDEN

Keyword(s): IEEE INSTRUMENTATION AND MEASUREMENT TECHNOLOGY CONFERENCE - SENSING, PROCESSING, NETWORKING (IMTC/97); OTTAWA, CANADA. MAY 19-21, 1997

Zurück | Weiter