Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Clustering-based pattern recognition applied to chemical recognition using SAW array signals

Conference ·
OSTI ID:10151490
We present a new patter recognition (PR) technique for chemical identification using arrays of microsensors. The technique relies on a new empirical approach to k-dimensional cluster analysis which incorporates measured human visual perceptions of difficult 2- dimensional clusters. The method can handle nonlinear SAW array data, detects both unexpected (outlier) and unreliable array responses, and has no user-adjustable parameters. We use this technique to guide the development of arrays of thin-film-coated SAW (Surface Acoustic Wave) devices that produce optimal PR performance for distinguishing a variety of volatile organic compounds, organophosphonates and water.
Research Organization:
Sandia National Labs., Albuquerque, NM (United States)
Sponsoring Organization:
USDOE, Washington, DC (United States)
DOE Contract Number:
AC04-94AL85000
OSTI ID:
10151490
Report Number(s):
SAND--94-0057C; CONF-9406166--2; ON: DE94012070; BR: GB0103012
Country of Publication:
United States
Language:
English