An Experimentally and Numerically Trained Neural Network Co

The autonomous control of a hypersonic missile from launch to target is a very challenging problem. F&S proposed to use a combination of experiments using MEMS and numeric simulations using CFD, to train a neural net for the control of micro-jets. The trained neural net control system would be able to control a hypersonic missile throughout its flight. Such a system could possibly be used to replaced probes in any aircraft, reducing drag, preserving stealth, and lower maintenance.An excellent development team has been assembled including Professor Joseph Schetz of Virginia Tech, a world leader in experimental high-speed aerodynamic measurements, Professor Ken Shaw of Virginia Tech, an expert in the development and training of neural networks, AeroSoft, a world leader in CFD software development, and F&S, a leader in fiber optic sensor development and commercialization. The F&S team is both qualified and motivated to build upon their combined demonstrated capabilities to accomplish this development and commercialization project.