机器人模糊神经网络阻抗控制

Fuzzy Neural Networks Impendence Control of Robot

提出一种以位置控制为基础的模糊神经网络阻抗控制结构,它根据力误差修正参考位置,使机器人与环境的实际接触力跟踪期望力。利用模糊神经网络学习参考位置,不需要环境位置和刚度的先验知识。这种方法有误差补偿作用,对干扰和环境等不确定因素具有鲁棒性,仿真结果表明了控制方案的有效性。

A position-based impendence control structure using fuzzy neural networks is proposed in this paper. It modifies the reference position according to the force error so that contact force between the manipulator and the environment can track the reference force. The fuzzy neural networks learn about the reference position so that the prior knowledge of the environment stiffness and location is not required. This method can compensate the errors and has robusticity to the uncertainties and disturbance of the environment. Simulation results show that this control scheme is effective.