THE ELECTRONIC NOSE - PowerPoint PPT Presentation

1 / 55
About This Presentation
Title:

THE ELECTRONIC NOSE

Description:

An Electronic Nose is a system that uses the pattern of ... Determine the odor. How to Work LabVIEW. Enter the file name. Enter Sample Code. Choose Autoscaling ... – PowerPoint PPT presentation

Number of Views:12649
Avg rating:3.0/5.0
Slides: 56
Provided by: iit61
Category:
Tags: electronic | nose | the | odor

less

Transcript and Presenter's Notes

Title: THE ELECTRONIC NOSE


1
THE ELECTRONIC NOSE ITS POTENTIAL
Miss Asha Joseph Dr. Peter Lykos
2
All Those Responsible
  • Dr. Peter Lykos
  • Dr. Joseph Stetter
  • Dr. Bill Penrose
  • Dr. Pravin Patel
  • Chris Morong
  • Asha Joseph
  • Yan Hu
  • John Carr
  • Luke Jarymowycz
  • Doug Sievers
  • Archan Mehta
  • Justin Henry
  • Jamal Hussein
  • Joel Ziemke
  • Jerome Hannecart
  • Zaki Shaikh

3
Contents
  • What is an Electronic Nose?
  • Comparison with a Biological Nose
  • Uses
  • Design of the Electronic Nose
  • Sampling and Distribution
  • Sensors and Electronics
  • LabVIEW and Pattern Recognition
  • The Experiment
  • Future of the Electronic Nose

4
What is an Electronic Nose?
  • An Electronic Nose is a system that uses the
    pattern of responses from an array of gas sensors
    to examine and identify a gaseous sample.

5
Our Electronic Nose
6
An Electronic Nose
7
Analogy Between the Electronic Nose and the
Biological Nose
  • Inhaling ? Pump
  • Mucus ? Filter
  • Olfactory Epithelium ? Sensors
  • Binding With Proteins ? Interaction
  • Enzymatic Reactions ? Reaction
  • Cell Membrane Depolarized ? Signal
  • Nerve Impulses ? Circuitry Neural Network

8
How Does a Biological Nose Work?
A ROSE !
Circuitry NN
Mucus (filter)
Receptor Protein (sensor)
Nerve
Would a rose smelled by any other nose smell as
sweet?
Input
9
The Basic Design of an Electronic Nose
Array of Signals
Vapor
Result
Pattern Recognition
Array of Sensors
Sample
10
Why is an Electronic Nose Better?
  • Trained human sniffers are expensive
  • Individuals vary
  • Hazardous Chemicals
  • Can be done in real time for long periods

11
Where Electronic Noses Are Used
  • Automotive
  • Food safety
  • Telemedicine
  • Emergency response
  • Military and space
  • Environmental monitoring

12
Nose Accessories
  • Computer Control
  • Modulation Sensor
  • Switches
  • Identical Sampling
  • Improved Pattern Recognition Software
  • Sampling System
  • Gas Collector

13
SENSOR MANIFOLD
14
First Design Consideration
Linear serial
  • Advantages
  • Easy to make
  • Small volume
  • Disadvantages
  • Degrades the sample
  • Cannot be greatly expanded

15
Second Design Consideration
Radial Parallel
Advantages Sample flows equally to each
sensor Disadvantages Complex to make Larger
volume than the linear manifold
16
Radial Sensor Manifold
17
Sampling What is a sniff?
18
What is Sampling?
  • Preparing the sample
  • Volatile liquids
  • Vapor phase sampled directly
  • Solids
  • Sample may be ground
  • Chemical extraction
  • Sample delivery
  • - Control Variables
  • Concentration
  • Pressure
  • Temperature
  • Volume
  • Flow rate

19
Sampling System
20
Switches and Filament
Controlled by LabVIEW Why Computer
Controlled Identical Sample Runs Easy Comparisons
21
Sampling System
22
Sampling System
23
Electronics Instrumentation
24
The Modulation Sensor (Filament)
  • Hydrocarbons do not cause significant response in
    electrochemical sensors
  • Acts like a pseudo-mass spectrometer
  • Wider Chemical Profile
  • Inactive molecules broken down into
    electrochemically reactive fragments

OFF ON
25
Filament controller circuit
26
Automation of the Sampling System
27
Electrochemical Sensors
GAS
Porous Membrane
Filter
Working Electrode
Gain
Reference
Common
Bias
Signal

H Electrolyte
28
CO Oxidation and Reduction
Oxidation Working Electrode CO H2O
CO2 2H 2e- Reduction Common Electrode O2
2H 2e- H2O CO
29
Schematic
30
LabVIEW Principal Component Analysis
31
What is LabVIEW?
  • A graphical programming language
  • Data Acquisition Techniques
  • Data manipulation, creation, and display
  • Control hardware

32
LabVIEW Block Diagram
33
What Can You Do With a Neural Net?
  • Just about anything that a regular computer can
    do, only better
  • Instead of calculating one step at a time,
    calculates many things at the same time
    (resembles parallel processing)

34
Common Neural Net Design
35
What is Principal Component Analysis?
  • Principal Component Analysis is a method of
    projecting higher dimensional space onto a two
    dimensional plane for easy visualization.
  • More of a Statistical Method

36
How to Perform PCA?
x X
  • Normalize Sensor Array

_
  • Scale Data (x-x)/s
  • Store all Data as an m x n matrix (A)

T
  • A A produces a n x n matrix
  • Take Eigenvalues (?) and Eigenvectors (V)

?
k
  • Variance

k
37
How to Perform PCA? (Cont..)
  • Use the two largest eigenvalues and find the
    corresponding eigenvectors
  • Take the two eigenvectors and dot them with each
    scaled sensor array (V A )?X
  • (V A )?Y
  • Plot (X,Y)

i
1
i
2
38
How Neural Nets Will Benefit the Electronic Nose
  • Sensor responses drift over time
  • Change with concentration
  • Sample Depletion
  • Relate the Pattern of the data
  • Find the data that is the closest
  • Determine the odor

39
How to Work LabVIEW
  • Enter the file name
  • Enter Sample Code
  • Choose Autoscaling
  • Click Acquire Data ON
  • Click the run Arrow
  • Collect Data
  • View the PCA

40
LabVIEW for Dummiees
http//www.iit.edu/labview
41
But Does It Work?
42
The Experiment
  • Five Samples
  • Broth
  • Escherichia coli
  • Pseudomonas aeruginosa
  • Diluted 51 with saline solution
  • Continuous stirring
  • Each Run 6 minutes
  • Repeated 7 times in succession
  • Tubes Flushed with air for 4-5 minutes
  • Next Sample

Staphylococcus aureus Enterococcus faecalis
43
  • Broth Run 6 (9-10-99)

44
  • Broth Run 6 (9-10-99)

45
PCA 9-10-99 Different Bacteria
46
MOSES II Coffee Data (PCA)
C3
U1
C1
C2
U2
X-Axis ?80 Variance (PC1) Y-Axis ?8 Variance
(PC2)
47
FUTURE OF THE E-NOSE
48
Future of the Electronic Nose
  • Research is being done on IC E-Noses
  • Miniaturizing Current Technology
  • Combination of several sensors
  • Conducting Polymer
  • Quartz Crystal Microbalances
  • Variable Capacitor
  • Manufacture of olfactory nerves

49
How and Why Responses Happen
  • Sensors respond to different functional groups
  • There is cross sensitivity and selectivity
  • Not all sensors respond the same way

50
Types of Sensors
  • Calorimetric
  • Conducting (Metal Oxide and Polymer)
  • Piezoelectric (QCM and SAW)
  • Electrochemical
  • NOTE All need transducers

51
Calorimetric
  • The sensors measure the concentration of
    combustibles species by detecting the temperature
    rise resulting from the oxidation process on a
    catalytic element.

52
Conducting Metal Oxide Polymer
  • The gas is absorbed onto the surface and the
    resistance changes
  • Voltage is proportional to the Resistance
  • The change in voltage is measured
  • Metal Oxide most popular (200-400C)
  • Polymer (Room Temp)

53
Piezoelectric QCM SAW
  • Quartz Crystal Microbalances
  • Quartz crystal resonates
  • Polymer coating adsorbs the gas
  • Mass change causes a change in frequency
  • Surface Acoustic Wave
  • Same principles
  • Measures the phase shift in the wave

54
Special Thanks
  • Dean Kallend - Monetary Support
  • Dr. Lykos, Dr. Stetter Dr. Penrose
  • Mentoring
  • Technical support
  • Thomas Torres - Machine Shop
  • Charles Wenger - Research Librarian
  • Citytech - Free Sensors
  • Bacharach Inc. - Free Sensors
  • Sensor Tek - Free Sensors
  • TSI Inc. - Free Sensors
  • Starbucks - Free Coffee Samples
  • Dr. Pravin Patel
  • Bacteria Samples

55
Questions?
Write a Comment
User Comments (0)
About PowerShow.com