On-Line Trajectory Optimization for Autonomous Air Vehicles

"Successful operation of next-generation unmanned air vehicles will demand a high level of autonomy. Autonomous low-level operation in a high-threat environment dictates a need for on-board, robust, reliable and efficient trajectory optimization. Theproposed effort will develop and demonstrate an innovative combination of traditional analytical and numerical solution procedures to produce efficient, robust and reliable means for nonlinear flight path optimization in the presence of time-varyingobstacles and threats. The solution procedure exploits the natural time-scale separation that exists in the aircraft dynamics using singular perturbation theory. A reduced order problem involving only the kinematics of the position subspace will betreated numerically. The nonlinear aircraft dynamics will be treated analytically in a boundary layer analysis that results in an optimal feedback guidance solution. The developed algorithms will be coupled with a neural network adaptive autopilot andintegrated in an existing unmanned testbed. Phase I will produce a demonstration of developed algorithm performance in near-real-time flight simulation. Phase II will expand the scope of mission scenarios that can be addressed, will produce flight codevalidated in real-time hardware-in-the-loop simulation, and culminate in close-range flight demonstration of the developed technology on a fixed-wing unmanned aircraft testbed. It is anticipated that the subject technology will find application in the widevariety of unmanned flight systems now under development. The technology is enabling for next-generation unmanned Department of Defense flight and weapon systems that will be required to operate autonomously at low levels in a high-threat environment.The technology can also be applied to civil and commercial unmanned systems operating in an urban setting, and can potentially be used with other evolving technologies to provide a much needed see-and-avoid capability for all unmanned flight systems."