直升机神经网络反馈线化飞行控制

以SH－2G为控制对象，采用基于神经网络的非线性系统反馈性化控制方法，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逢适应能力，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。

Flight Control of Helicopter by Feedback Linearization Based on Neural Networks

A maneuver simulation of the helicopter, modeling of SH-2G, is performed by using the nonlinear system control technology for feedback linearization based on neural network. Simulational result shows that the neural network is adaptive. It has the online capability of compensating the inverse error caused by feedback linearization. This approach can offer effective control throughout the full flight envelope and only the static inverse model based on a single flight condition is needed, instead of the dynamic inverse model. It solves some problems of gain scheduled, for using conventional approaches, difficulty for constructing the dynamic inverse model and intensive online computation. So it is a kind of intelligent control technology and has great potentialities.