Nonlinear Least-Squares Based Method for Identifying and Quantifying Single and Mixed Contaminants in Air with an Electronic Nose

Abstract

The Jet Propulsion Laboratory has recently developed and built an electronic nose(ENose) using a polymer-carbon composite sensing array. This ENose is designed to be usedfor air quality monitoring in an enclosed space, and is designed to detect, identify andquantify common contaminants at concentrations in the parts-per-million range. Itscapabilities were demonstrated in an experiment aboard the National Aeronautics and SpaceAdministration’s Space Shuttle Flight STS-95. This paper describes a modified nonlinearleast-squares based algorithm developed to analyze data taken by the ENose, and itsperformance for the identification and quantification of single gases and binary mixtures oftwelve target analytes in clean air. Results from laboratory-controlled events demonstrate theeffectiveness of the algorithm to identify and quantify a gas event if concentration exceedsthe ENose detection threshold. Results from the flight test demonstrate that the algorithmcorrectly identifies and quantifies all registered events (planned or unplanned, as singles ormixtures) with no false positives and no inconsistencies with the logged events and theindependent analysis of air samples.

Keywords

• electronic nose
• sensor array data analysis
• nonlinear least squares.