基于RBF神经网络的全向智能轮椅自适应控制

Adaptive control for omni-directional intelligent wheelchairs based on RBF neural network

在外界扰动为有界不可测条件下,利用径向基函数(radial basis function,RBF)神经网络在线逼近全向智能轮椅的非线性逆运动学模型,提出对轮椅轨迹跟踪的直接自适应控制方法.首先,在分析全向智能轮椅平台动力学模型的基础上,设计了基于径向基函数神经网络的全向智能轮椅自适应控制器;并进一步利用李雅普诺夫稳定性理论,证明了在外界扰动及神经网络权值误差逼近有界的条件下,该控制器在全向智能轮椅轨迹控制中跟踪误差的一致稳定且有界;最后,通过全向智能轮椅轨迹跟踪仿真实验,验证了所提出控制方法的有效性和稳定性.

Under the condition that the disturbance was bounded and immeasurable, a direct adaptive control method based on radial basis function （RBF） neural network for omni-directional intelligent wheelchairs was proposed. Firstly, on the basis of analyzing inverse-dynamic model of intelligent wheelchair platform, an adaptive controller was proposed based on RBF neural network for omni-di- rectional intelligent wheelchair. Furthermore, by using Lyapunov stability theory, it had been proved that the tracking error of this controller was stable and bounded on condition that the disturbance and network weight were bounded during the intelligent wheelchair＇s control process. Finally, the validity and stability of the proposed method were verified through the simulations of the omni-directional in- telligent wheelchair.