Prototype surface acoustic wave chemical sensor systems are described, which can detect and identify toxic vapors in real-time at trace concentrations. To operate autonomously for long periods, without failure, requires a thorough understanding of the hardware and software requirements of the sensor system. The SAWCAD and SAWRHINO prototypes, which implement several improvements to the hardware, over previously developed systems, are described. Software for vapor detection and neural network identification are also discussed. Preliminary results from two new software enhancements are described. Improved chemical discrimination occurs when the response slopes are incorporated into the analysis of the SAW ambient data. The generalized rank annihilation method is shown to be a powerful tool for extracting pure component analyte signatures from trap and purge gas solid chromatographic SAW data.
|