论文部分内容阅读
以 SH-2 G为控制对象 ,采用基于神经网络的非线性系统反馈线性化控制方法 ,进行直升机机动飞行仿真。仿真结果表明 ,所设计的神经网络具有自适应能力 ,能够在线补偿反馈线性化所产生的逆变换误差。这种方法无须获得动态逆模型 ,而只需某一个状态下的动力学逆模型 ,就能在全包线提供有效的控制 ;既避免了传统方法的增益调参 ,又解决了难以获得动态逆模型及计算量大的问题 ,是一种很有发展潜力的智能控制方法。
Taking SH-2 G as the control object, a nonlinear system feedback linearization control method based on neural network was used to simulate helicopter maneuver flight. The simulation results show that the designed neural network has adaptive capability and can compensate the inverse transform error generated by feedback linearization online. This method need not obtain the dynamic inverse model, but only need a certain state inverse kinematics model, can provide effective control in the whole envelope; not only avoid the traditional method of gain parameters, but also to solve difficult to obtain dynamic inverse The problem of large model and calculation is a kind of intelligent control method with great potential for development.