SBIR-STTR Award

The objective of this proposal is to develop a reliable method to obtain reduced-order models for the simulation of large-scale seismic and acoustic signature propagation. These reduced-order models should be capable of reproducing results comparable to those obtained from a high-fidelity high-performance geologically complex numerical simulation code that normally requires a massively parallel computer to run. Furthermore, these reduced-order mathematical models will be sufficient in providing decision-aid tools for choosing sensor locations and for predicting sensor performance.

Keywords:
Reduced-Order Models, Seismic, Acoustic, Signature Propagation, Sensor Placement

In Phase 1 research, we developed algorithms to produce reduced-order models from High Performance Computations (HPC). These algorithms are unique, powerful, and particularly well-suited to the acoustic and seismic signal propagation problem. In this Phase 2 proposal, we seek to extend these algorithms and produce HPC plug-ins that can be integrated with existing HPC codes. Not limited to acoustic or seismic signal propagation, these plug-ins can also be used in other HPC applications where the relevant dynamics are linear or can be linearized, e.g., antenna and propagation models. We will develop effective sensor placement algorithms to monitor a targeted area. We will also produce algorithms for fast identification of source location and source signal for anti-sniper application. These algorithms will be tested on research-level acoustic models, and on operation-level HPC model of the McKenna MOUT site at Fort Benning. The work to be performed in this Phase 2 project will lay a firm foundation for Phase 3 work where we will conduct field experiments and develop complete packages with integrated hardware and software for both military and commercial applications.

Keywords:
Performance Computing (Hpc), Reduced-Order Models, Signature Propogation, Acoustics, Source Localiza