Design, implementation, and field testing of a portable fluorescence-based vapor sensor.

The design and implementation of a portable fluorescence-based vapor sensing system are described. The system incorporates previously developed microsensor array technology into a compact, low-power device capable of collecting and delivering ambient vapor samples to the array while monitoring and recording the fluorescent responses of the sensors. The sensors respond differentially when exposed to a sample vapor and, when processed using a support vector machine (SVM) pattern recognition algorithm, are shown to discriminate between three classes of petroleum distillates. The system was characterized using sample vapors prepared under several different conditions in three sensing scenarios. The first scenario demonstrates the basic operational capability of the device in the field by presenting high concentration vapors to the array. The second scenario introduces the potential for a greater degree of variability in both sample vapor concentration and composition in an effort to emulate real-world sensing conditions. The third scenario uses an on-board trained pattern recognition algorithm to identify unknown vapors as their responses are collected. The device demonstrated high classification accuracy throughout the field tests and is capable of improving its classification accuracy when challenged with samples presented under variable ambient conditions by enhancing the signal-to-noise ratio of the array response.