Tetlow, M.Evans, M.Schneider, G.Schottle, U.2008-05-092008-05-09200543rd AIAA Aerospace Sciences Meeting and Exhibit [Proceedings], 2005, pp.AIAA 2005-819-1-AIAA 2005-819-129781624100642http://hdl.handle.net/2440/43652The aim of this study is to develop and test a robust guidance strategy for use on the flyback mission of a winged launch booster. The guidance system employs a numerical Newton-Raphson restoration technique to update steering parameters, coupled with heuristic logic, to improve robustness. The guidance system uses a prediction of the final state of the launch vehicle to close the feedback loop and is therefore called “Predictive guidance”. Virtual environment and virtual vehicle models are used to simulate the booster and the environment in which it operates. The virtual booster passes the current velocity and position to the predictive guidance system at given guidance intervals. The predictive guidance system then integrates along the trajectory, using the current parameterized steering model, to determine the expected final position of the virtual booster. It compares the achieved final state to the required target state and calculates the target error. A parameterized non-linear restoration technique then determines new values for the steering parameters, to guide the virtual booster from the current state to the desired state. The final phase of the flyback mission is guided by a series of controllers, which achieve and maintain the required cruise flight conditions. Once the guidance system is operational for the nominal case, various errors are introduced in the virtual environment to test robustness. A Monte Carlo type analysis is performed to test the guidance system operation in 1000 different flight regimes, with randomly varied environmental and state parameters. The randomly varied parameters include: wind speed and direction, staging flight path angle, velocity and altitude and modelled state estimation errors. Different gravitation and atmosphere models are also used in the virtual environment and guidance computer to introduce further differences between the guidance estimate of the flight environment and the virtual flight environment. The guidance system is found to be fast enough for real-time implementation and able to achieve cruise conditions in all cases. The Monte Carlo analysis also allows for a statistical analysis to determine the mean and standard deviations for the target conditions.enCopyright © 2005 by Matthew Tetlow. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.Conference paper00200773742008050916575244261