一种柔性机器人的神经网络控制方法

提出了一种末杆为柔性杆的机器人的新型神经网络控制方法.它将末杆为柔性杆的机器人分解成刚性前端和柔性末端两部分,使用传统的机器人控制方法控制刚性前端,采用神经网络辩识柔性末端部分的逆模型,将末端的期望轨迹实时分解成末关节转角及末关节轴位姿的期望轨迹再分别进行控制.这一方法将传统的基于模型的控制方法和神经网络方法有机结合在一起.通过对平面3杆柔性机器人的数值仿真,初步验证了这一方法的有效性和良好效果.

Novel Neural Networks Control Approach of Flexible Manipulators

A novel neural networks control approach for manipulator with flexible end link was proposed. The manipulator with flexible end link was regarded as two parts: the flexible latter part that includes the last joint and the last link, and the rigid former part that is composed of the rest elements. In the proposed approach, a conventional model based controller was used to control the rigid former part, a neural network was used to approximate the dynamic anti model of the flexible latter part. Combined both together, the proposed approach achieved fine control effect in the emulation of a three links planar manipulator with a flexible last link.