Reentry trajectory design for Reusable Booster Vehicles based on neural network

Reusable Booster Vehicles(RBV) is a new type of Two-Stage-To-Orbit(TSTO) partly reusable launch vehicle, which can greatly reduce launch cost. The guidance control during the reentry period of RBV is critical for safe and reliable return. In order to satisfy special task demands, this paper has analyzed the performance of neural network guidance, normal trajectory guidance and prediction-correction guidance, and given related comparison conclusions. The neural network has been trained according to the trajectory parameters generated by using prediction-correction guidance method. The state parameters of RBV have been used as the input, and the bank angle has been the output. Simulation result has indicated that the proposed method has good real time performance and its computation occupation is low. What´s more, it also has good robustness and adaption performance by using the proposed method. In the case of larger initial bias or greater error of RBV pneumatic parameters, it could attain good land precision. It has a very good application future by utilizing the reentry guidance method for RBV based on neural network.