带有神经网络干扰观测器的视线角约束制导

Aiming at the problem of maneuvering target interception with terminal line-of-sight (LOS) angle constraint, a LOS angle constraint guidance method based on radial basis function (RBF) neural network interference observer is proposed. Firstly, considering that the acceleration information cannot be obtained during target maneuvering process, an interference observer based on RBF neural network is presented, which realizes high-precision estimation of target maneuvering. Secondly, an improved sliding mold guidance law is designed by introducing the power term by fully considering the terminal angle constraint and combining the idea of super-twisting algorithm, so as to effectively improve the guidance accuracy under limited overload conditions. On this basis, the convergence and stability of the algorithm are proved by Lyapunov' s theorem. Finally, the guidance performance of three different methods in four interception scenarios is compared through simulation verification, and Monte Carlo simulation is given for the proposed method, and the simulation results show that the LOS angle constraint guidance law given in this paper has high accuracy and strong robustness for maneuvering target interception.