论文部分内容阅读
为了研究具有外部干扰力矩和模型不确定性的多体航天器姿态快速跟踪控制问题,基于逆系统方法和回声状态网络(echo state network,ESN),设计了鲁棒控制器,并利用Lyapunov稳定性理论证明了控制系统的渐近稳定性。采用“meta-learning”策略离线训练ESN网络,并应用遗传算法优化其主要参数,解决了动态递归神经网络训练困难及网络参数不易确定的问题。控制器的设计过程相对简单,不需要精确的动力学模型。该文还针对航天器中心体与天线同时跟踪不同目标的任务进行了数值仿真,结果表明所设计的鲁棒控制器对外部干扰与模型不确定具有很好的鲁棒性。
In order to study the fast tracking control of multi-body spacecraft with external disturbance torque and model uncertainty, a robust controller is designed based on the inverse system method and the echo state network (ESN). The Lyapunov stability The theory proves the asymptotic stability of the control system. The “meta-learning” strategy is used to train the ESN network offline. The genetic algorithm is used to optimize the main parameters of the ESN. This solves the problem that the training of the dynamic recurrent neural network is difficult and the network parameters are not easy to determine. The design process of the controller is relatively simple and does not require an accurate kinematic model. The paper also simulates the task of the spacecraft body and the antenna to track different targets at the same time. The results show that the designed robust controller has good robustness against external disturbances and uncertain models.