论文部分内容阅读
选择具有常值加速度的目标作为一个简化但有代表性的目标模型，以拦截弹的Δｖ最小为性能指标，得到了一定初始拦截距离下的训练数据，并使用ＢＰ算法对神经网络进行了训练。在有目标机动加速度估计误差的情况下，将神经网络制导获得的拦截精度和变轨机动速度增量同用扩展比例导引得到的结果进行了比较。结果表明，神经网络导引法能够保证对目标的拦截精度，具有很强的适应性。
Select a target with constant acceleration as a simplified but representative target model, and obtain the training data under a certain initial interception distance with the minimum Δv of the interception missile as the performance index. The neural network is trained by the BP algorithm. In the case of the target maneuvering acceleration estimation error, the intercept precision and the incremental orbit maneuvering speed gained by the guidance of neural network are compared with the results of the extended proportional guidance. The results show that the neural network guidance method can ensure the accuracy of the target interception, and has strong adaptability.