The Study Of Glider Vehicle’s Predictive Guidance Based On NN Method

The traditional guidance was hard to resolve the contradiction between robustness,real-time and the guidance of precision. The research been to provide technical basis forhypersonic sliding vehicles with high precision of strike and availability of breakingthrough defenses, by combining neural network with forecasting guidance. Thisresearch about the new style guidance of sliding vehicles was exploration. The mainresearch contents and results of the paper are as follows:For the different combining fashion of neural networks and predictive guidance,the Error of Target Optimize Nero-Network (ETO-NN) and the Pattern of GuidanceController Study Nero-Network (PGS-NN) forecasting guidance were put forwardinnovatively, and the designed method and the full frame of these forecasting guidncewere studied, and the characteristic of these guidance was analysed in the research.The research of quickly and automatically obtaining the sample data with a certainquality was done, by designing the standard trajectory and the optimal prective guidance,and optimizing the iterative process of the optimal guidance algorithm. A3-layer BPneural network controller was designed, based on the changes of online flightparameters and environment, and the study of segmented scheduling was studied. Theoptimal prective guidance was studied as the pattern, samples producing and neuralnetwork training was researched, the design and assessee of PGS-NN prective guidancewas done, and the engineering application of PGS-NN forcasting guidance wasinnovatively researched in this paper.Finally, a whole process simulation study of the optimal forecasting guidancemethod and neural network predictive guidance method was carried out, and the resultsof two guidance methods were compared and analyzed. The results show that the neuralnetwork predictive guidance method has the expansive application future, and with nicereal-time and strong robustness.