Transition of Environmental Sampling Research to Operational Collections

Research and development for environmental sampling have often focused on deployed size and power reductions though improvements in sample media in order to reduce footprint and sample volume; however, the number of media samples that must be collected and stored in the field and consequently returned to the laboratory for analysis is a significant barrier impacting operationalizing technology. Source characterization from environmental samples using deployed collection systems faces a realistic expectation that positive sampling periods will have a sparse representation in the total collection space where overlap between signal, favorable transport conditions, and sampler location is likely to be small. Physical limits and processing costs indicate that increasing sampling for characterization may require more samples than practical to acquire due to size weight and power limits, or become too costly to analyze, while a significant fraction of analyses will be null or background detections. Compressed sensing strategies currently applied in digital signal processing and advanced imagery applications enable signal reconstruction from linear and/or non-linear optimization of very few nonadaptive measurements. Application of algorithms suited to underdetermined matrices can reduce the number of samples required for extended environmental monitoring and streamline subsequent data analyses by reducing the number of null/background-only samples in a sparse environment. An evaluation of compressed sensing strategies enabling signal reconstruction from linear and non-linear optimization of very few measurements enabling greatly improved analytical throughput will be presented.