基于人工神经网络的直升机智能飞行控制系统

Intelligent Helicopter Flight Control System Based on Artificial Neural Network

针对当前飞行控制系统难以在恶劣天气或者系统发生故障等紧急情况下完成自动飞行的局限性,笔者提出一种基于人工神经网络的直升机智能飞行控制系统。该系统模仿和学习大量人类飞行员的操作,建立起飞行数据和飞行员控制指令的数据映射,对人工神经网络进行离线训练,从而得到泛化的控制系统模型,在实际使用时输入实时飞行数据,人工神经网络在线计算操纵指令以控制直升机飞行。仿真结果表明:系统具有模拟人类飞行员自主控制直升机并保持稳定飞行的能力,在扰动条件下能够实现对直升机的鲁棒稳定控制,具有比人工操作更好的控制效果。

In view of the limitation that the current flight control system is difficult to complete automatic flight in severe weather or system failure and other emergency situations,the author proposes an artificial neural network based helicopter intelligent flight control system.The system imitates and learns a large number of human pilots'operations,establishes the data mapping of flight data and pilot control instructions,and conducts off-line training for the artificial neural network,so as to obtain the generalized control system model,input the real-time flight data in actual use,and the artificial neural network calculates the control instructions online to control the helicopter flight.The simulation results show that the system has the ability to simulate the human pilot's autonomous control of the helicopter and maintain stable flight.Under the disturbance condition,it can achieve robust stability control of the helicopter and has better control effect than the manual operation.