自由漂浮空间机械臂基于神经网络的鲁棒自适应控制

考虑数学模型难以精确获得及带外部干扰情况下，针对自由漂浮空间机械臂的轨迹跟踪控制问题，提出一种基于神经网络的自适应鲁棒控制策略。基于Lyapunov稳定性理论设计理想控制器，进而推出系统的不确定模型。利用神经网络的学习能力逼近系统不确定模型，从而避免保守上界的估计。利用线性化技术并结合Lyapunov函数，设计包括权值及隐层参数在内的在线自适应学习律及鲁棒控制器，加快了误差收敛速度及控制精度，并消除了高阶逼近误差及扰动，保证了系统的一致最终有界，仿真比较表明了该控制策略的有效性。

Robust Adaptive Control for Free-floating Space Manipulators Based on Neural Network

Trajectory tracking of a class of free-floating space manipulators with disturbance and model uncertainties are considered. Robust adaptive control scheme based on neural network is proposed. Ideal controller according to the Lyapunov stability theories is designed. Uncertainty model is derived. Neutral network is used to adaptive learn and compensate the unknown system, the estimate of conservative boundary is avoided. Linear parametric technology combining Lyapunov function is used to design the on-line real time adjust learning law including weights value and network parameters. The scheme accelerates the convergence velocity and control accuracy and eliminates high level to approach errors and perturbation. The controller can guarantee uniformly ultimately bounded. The simulation results show that the presented methods are effective.