Volatile Organic Compound Odor Signature Modeling

ABSTRACT: Continuing the Phase I research involving VOC Odor Signature modeling, Toyon Research Corporation and North Carolina State University proposes to develop two products of the Phase II effort: VOCScenario and VOCSourceSeek. VOCScenario will provide a simulation environment for supporting the development emerging e-nose sensors for Volatile Organic Chemicals (VOCs). The solution relies on the ability to predict the spatio-temporal distribution of signatures formed by mixtures of VOCs at the appropriate resolution necessary to support further studies, over a range of applications that includes indoor and outdoor environments, and a variety of sources. Features include data visualization, pre-defined scenarios, and a sensor modeling capability to emaulate actual sensor results. The architecture supports honest evaluation of source localization algorithms. VOCSourceSeek is intended to quickly identify the source using handheld sensors. The envisioned solution relies on localization algorithms that use a mixture of environmental knowledge and real-time sensor data to estimate the position and strength of a source. The Phase II effort will consider two proposed algorithms: one based on a gradient-based search, and the other using Bayesian inference techniques. The Phase II effort involves assessing each algorithm in simulation and with a Proof-of-Concept test that implements a prototype VOCSourceSeek concept. BENEFIT: VOCScenario provides AFRL with a robust means of establishing sensor requirements investigating new ConOps, assessing localization algorithms, and determining the time and length scales of interest for a wide spectrum of applications. The results will mitigate the technical risk associated with new sensing technologies by enabling researchers to conduct trade studies and analyses with a synthetic dataset that has much more fidelity and resolution than could ever be affordably measured in a test environment. Potential uses for the localization architecture and algorithms that are part of VOCSourceSeek can be found in many applications, including leak detection, explosive detection, victim search and rescue, tracking individuals, and indoor air quality surveys.