Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Chemical analysis
Reexamination Certificate
2000-07-18
2002-06-25
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Chemical analysis
C073S023340, C073S031050, C073S031060
Reexamination Certificate
active
06411905
ABSTRACT:
FIELD
The present invention relates to a method and apparatus for obtaining useful measurements from an electronic nose and, more particularly, to a method and system for obtaining an estimate of odor concentration from data obtained from a sensor-array type electronic nose.
BACKGROUND
Odor concentration is measured by determining the dilution factor required to reach the human detection threshold. The odor concentration is then expressed in terms of multiples of the detection threshold. The detection threshold is defined as the dilution factor at which the sample has a 50% probability of being detected under the conditions of the test.
Odor concentration is conventionally measured by an instrument called an olfactometer, the use of which requires a panel of persons who have been selected and trained following standard methods (see ASTM Special Technical Publication 758: Guidelines for Selection and Training of Sensory Panel Members; CEN/TC264/WG2 1998: Draft European CEN Standard: Air-quality-determination of odour concentration by dynamic olfactometry). Using an olfactometer with a human panel is presently the most precise method for quantifying odors. However, the use of human panels to evaluate odor samples is labor intensive, time consuming, prone to errors, and cannot be used a remote sites. The measurement accuracy obtainable with an olfactometer is dependent upon both the olfactometer and the human panel. Huge variations in olfactory sensitivity are found, even among “normal people”. There is a clear need for a less labor-intensive, mobile way to measure odor concentration that is at least as accurate as an olfactometer.
Electronic noses, such as the AromaScan™ electronic nose (commercially available from AromaScan Inc., 14 Clinton Drive, Hollis, N. H.), use conducting polymer sensor arrays to mimic the human olfactory system in the classification, discrimination, and recognition of chemical patterns occurring in odor samples. However, electronic noses have not been successfully used to make quantitative measurements of odor concentration as defined above, but have been used to make measurements of patterns of odor mixtures and of physical parameters such as partial pressure and mole fraction. While such measurements are useful in quality control in food, beverage, perfume, and cosmetic industries and in detecting dangerous concentrations of pollutants, they are not useful in determining whether an odor will be detected by humans and quantitatively how offensive an odor will be to humans.
An electronic nose works by measuring the changes in electrical resistance of the sensors when exposed to an odor. The AromaScan™ electronic nose, for example, has 32 different sensors in its array, each of which will in general exhibit a specific change in electrical resistance when exposed to air containing an odor. The selective interaction of odors with the sensors produces a pattern of resistance changes for each odor. If an odor is composed of many chemicals, the pattern will be the result of their combined interactions with all of the sensors in the array. It has also been found that the response of the array to varying concentrations of the same odor is non-linear.
There have been some suggestions in the research literature that an electronic nose could be used for measurement of odor concentration. For example, see Hobbs et al. (Assessment of odours from livestock wastes by a photoionisation detector, an electronic nose, olfactometry and gas-chromatography-mass spectrometry. J. Agric.Engng Res. (1995) 60, 137-144) in which an evaluation of the response of an electronic nose against odor concentration measurement by an olfactometer is described. The electronic nose used was not sufficiently sensitive to be used to measure concentration, but could distinguish types of odors at high concentrations. There was no discussion of how odor concentration might be esitmated from the sensor-array response as there was no range of odor concentration over which the sensor array would respond.
By the time of a paper by Persaud et al. (Assessment of conductng polymer odour sensors for agricultural malodour measurements. Chemical Senses 1996, 21, 495-505), sensor array construction appeared to have improved as in that paper the response of a sensor array to the odor of a number of components of pig slurry was observed over a range of concentrations. It was observed that the normalized responses of the sensors to the various components changed with concentration. Reduction of the number of dimensions in the data was used in the analysis. The effect of humidity was also observed, but a correlation between sensor sensitivity and human sensitivity was not determined, nor was a method suggested for doing so. However, use of conducting polymer sensor arrays was found suitable for odor concentration estimation.
Misselbrook et al. (Use of an Electronic Nose to Measure Odour Concentration Following Application of Cattle Slurry to Grassland J. Agric.Engng Res. (1997) 66, 213-220) describes work done with two types of electronic nose, one of which was an AromaScan™ electronic nose. The paper describes calculation of normalized sensor responses and the application of principal components analysis to sets of normalized sensor responses at different odor concentrations and to sets of actual sensor readings. The analysis reported in the paper indicated that there was a definite concentration effect using in sets of actual sensor readings, although no regression (linear or otherwise) was done. The paper suggests that “using pretrained knowledge about the relationship between sensor response (either average sensor response or actual sensor response pattern) and odor concentration” for a particular odor type, it should be possible for an electronic nose to estimate a value for the odor concentration of a sample with which it is presented. This paper does not describe or suggest how to use neural networks to find a relationship between sensor response and odor concentration.
The only patents known to the inventors in which neural networks are mentioned in connection with measurement of odors by sensor arrays are those in the family of U.S. patents assigned to the California Institute of Technology (U.S. Pat. Nos. 5,571,401; 5,698,089; 5,778,833; 5,891,398; 5,911,872; 5,951,846; 5,959,191; 6,010,616; and 6,017,440). However, these patents do not describe at all how to use a neural network to obtain an estimate of odor concentration. They do disclose use of a sensor array and the use of principal components analysis to reduce the dimensionality of the data obtained from a sensor array. However, they do not describe or suggest the use of a neural network or any other non-linear fitting technique to convert preprocessed sensor array data to an estimate of odor concentration. In fact, no reference is made at all to estimating odor concentration or providing an electronic substitute for an olfactometer. The fitting that they describe is a multi-linear least squares fitting to obtain physical parameters such as partial pressure and mole fraction, not odor concentration. In particular, they describe using a multi-linear least squares fit through the first three principal components to the determine mole fractions of components in a mixture of gases.
No other patents known to the inventors describe a method or an apparatus that could be used for estimating odor concentration by a single direct measurement by a sensor array, although two patents (Hayashi, U.S. Pat. Nos. 5,627,307 and 6,006,583) describe apparatus for measuring odor concentration by repeated dilution down to a level at which a sensor produces a voltage known to indicate that the detection threshold for humans has been reached. Neural networks have been described in conjunction with devices for identification or discrimination of odors in several patents, but none suggested use in conjunction with odor concentration estimation.
There is thus a need to develop a method and apparatus for using an electronic nose to estimate odor concentration.
SUMMARY
In on
Armstrong William W.
Coleman Richard N.
Feddes John R.
Guoliang Qu
Leonard Jerry J.
Hoff Marc S.
Suarez Felix
The Governors of the University of Alberta
LandOfFree
Method and apparatus for estimating odor concentration using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for estimating odor concentration using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for estimating odor concentration using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2901460