柔性航天器自适应多层神经网络跟踪控制方法

Adaptive multilayer neural network tracking control for flexible spacecraft

针对柔性航天器的姿态跟踪以及振动抑制问题,提出了一种自适应多层神经网络控制方法。利用自适应多层神经网络来补偿系统的非线性项,利用光滑变结构项来补偿神经网络逼近误差及外部干扰。柔性航天器为典型的一个中心刚体加柔性附件的结构,假设模型参数未知并且具有任意的有限维。控制器只利用姿态角和角速度信息进行反馈控制,不需柔性附件振动信息。最后,实验表明该方法可以有效地完成姿态跟踪和振动抑制。

Aiming at the problem of attitude tracking and vibration suppression of a flexible spacecraft, an adaptive multilayer neural network （MNN） control method is proposed. The adaptive MNN is used to alleviate the nonlinearities and uncertainties of the system. Besides, a smoothed variable structure is applied to eliminate the approximate error and external disturbance of the MNN. The spacecraft consists typically of a rigid body and some flexible appendages, and it is assumed that the system parameters are unknown and the truncated model of the spacecraft has finite but arbitrary dimension. Only the attitude angle and its derivative are accessible for feedback control and the elastic modes are not needed. Finally, theory analysis and experimental results show that the attitude tracking and vibration suppression are accomplished effectively by the proposed method.