Reduced-Order High-Fidelity Models for Signature Propagation/WAVE

A reduced-order signature propagation model, with comparable full scale finite-difference time-domain (FDTD) accuracy is proposed. The reduced-order model achieves reductions in computational resources and time. Existing FDTD acoustic and seismic sensor models require supercomputing / multiprocessor level resources, which are impractical in a real time military / battlefield environment. The proposed effort combines efficient computational / numerical techniques with simplified physics and/or geometry to generate a reduced-order model to be evaluated against the benchmark FDTD model. Calculated impulse responses from the developed models, including a high order model FDTD, are compared using a least squares matching criteria. The accuracy of the localization of events is assessed using test cases accounting for sensor and source locations, and signal to noise levels. A real-time decision-aid tool provides a framework to perform trade studies, and generates a visual rendering illustrating results of the performance evaluations. Phase I focuses on development of an efficient reduced order code and well defined signal localization criteria working simultaneously within a visual environment to optimize the number and placement of seismic or acoustic sensors.