针对飞艇单发停车故障,提出了一种基于径向基神经网络的模型跟随非线性重构控制策略,并使用遗传算法对RBF网络进行了优化。该方法通过响应误差构造神经网络的学习规则,神经网络的输出驱动系统向着消除误差的方向运动。理论分析和仿真验证表明,该控制策略具有良好的重构性能和鲁棒性。 Aeroengine neural network model-following nonlinear reconfiguration control strategy is proposed to solve the single-engine parking problem. The genetic algorithm is used to optimize the RBF network. The method constructs neural network learning rules by responding to errors, and the output driving system of the neural network moves in the direction of eliminating error. Theoretical analysis and simulation results show that this control strategy has good reconstruction performance and robustness.