应用于力矩受限自由漂浮空间机器人的神经网络自适应控制

Neural Network Adaptive Control Applied to Free-Floating Space Robot with Bounded Torque

在控制力矩受限情况下,为实现具有模型不确定性自由漂浮空间机器人的轨迹跟踪控制,文章设计了一种新的神经网络自适应控制策略;首先,用双曲函数对控制力矩输入进行限制;其次,设计一种神经网络自适应控制律,对输入力矩受限条件下的非线性系统模型进行在线逼近,同时,利用鲁棒项对神经网络逼近误差和外界干扰进行消除;最后,根据李雅普诺夫理论,证明了所设计控制策略能够使自由漂浮空间机器人系统渐进稳定;仿真实验表明,该控制策略在无需建立复杂系统模型的情况下,便能够对控制力矩进行有效限制,从而使自由漂浮空间机器人在控制力矩受限情况下得到较好的控制。

In order to achieve the trajectory tracking of a class of free--floating space robot with bounded torque, disturbance and model uncertainties, a new control strategy is used in this paper. Firstly, using the hyperbolic function to restrict the control torque. Secondly, the control strategy design the neural networks adaptive law to approach the uncertain system model with bounded torque on--line, and the neural network approach errors and outside bounded disturbance can be eliminated by a robust controller. At last, the control strategy can guarantee the asymptotic convergence of tracking errors based on the Lyapunov stability theory, and can limit the range of control torque. The simulation results show that the control strategy can be free from depndence on the model, restrict the control torque and work effectively with bounded torque.