论文部分内容阅读
本文提出了一种利用神经网络逼近具有不确定性及随机干扰的仿射非线性系统的新算法，采用自适应控制率在线调节网络权值，并基于控制理论选择控制量以削减噪声干扰．从理论上证明采用该算法后系统全局稳定性．最后将该算法用于两连杆机械手轨迹跟踪，仿真结果表明该算法具有跟踪精度高，收敛速度快的优点．
In this paper, we propose a new algorithm for approximating affine nonlinear systems with uncertainties and stochastic disturbances by using neural networks. The network weights are adjusted online using adaptive control rates and the amount of control is chosen based on the control theory to reduce noise interference. It proves theoretically the global stability of the system after adopting the algorithm. Finally, the algorithm is applied to trajectory tracking of two-link manipulators. The simulation results show that the algorithm has the advantages of high tracking accuracy and fast convergence.