Abstract: The T-S model of the lateral mode of helicopter is developed based on the perturbation linear equations at typical flight states within the flight envelope.Then adopting parameter-mapping method and part of the states information，the qualifictory parameter sets are obtained according to given specifications，where the parameters of the controller are chosen.According to parallel distributed compensation algorithm，parameters at each typical flight states are employed to train a fuzzy-neural network to implement the control law，so that the helicopter can satisfy the specifications in the whole flight envelope.Simulations results indicate that，the control law designed above can assign system poles near the desirable ones in the flight envelope and that the performance of the system satisfies the specifications completely.Thus，the technique presented in this paper is feasible and effective.

Key words: T-S model, radial basis function neural network, parameter mapping, flight control

摘要： 根据飞行包线中典型状态的小扰动线性化方程，建立了直升机横侧向系统的T-S模型。采用参数映射设计方法，利用系统的部分状态信息，直接针对规定的系统性能指标，在满足要求的可用参数集中设计控制器参数。根据平行分布补偿原理，设计模糊神经网络控制器，利用各典型状态下选定的参数作为样本，训练模糊神经网络，实现所设计的控制律，使飞机能够在全包线范围内达到要求的性能指标。仿真结果表明，应用所设计的控制律，可以在飞行包线内将系统极点较准确地配置在希望极点附近，系统动态性能指标完全满足规定的要求。表明提出的设计方法可行而且有效。

关键词: T-S模型, 径向基函数神经网络, 参数空间设计, 飞行控制