Robust Algorithm Development for Multidimensional Chemical Analysis

Highly reliable detection of hazardous materials is a fundamental part of homeland security. Arrays of simple sensors can provide much of the sensitivity and selectivity of sophisticated sensors, but without the substantial hardware overhead. Unfortunately, such arrays are not without their challenges. The selectivity of such arrays can be realized only if the data is first distilled using advanced signal processing algorithms. There are several standard mathematical approaches that have been applied to this processing and while these techniques are certainly applicable to a subset of the problems of interest, they are not ubiquitous in their effectiveness. We choose to attack such problems using algorithms from the stochastic state estimation and data fusion regimes since they address precisely the same type of underlying problem. These algorithms promise to be enablers of advanced multi-dimensional chemical analysis, allowing a transition of complexity from the sensor suite to the data processing algorithms while still maintaining robust sensitivity and selectivity for analytes of interest; the ability to effectively use simpler and easier to fabricate sensor suites has the direct effect of placing lighter, more robust equipment into the hands of first responders.