基于模糊高斯基函数神经网络控制的机器人视觉伺服系统

Robotic Visual Servo System Based on Fuzzy Guass Function Neutral Network Control

针对机器人视觉伺服控制系统中存在的不确定性问题,将模糊控制与神经网络相结合,用神经网络来实现模糊推理,提出了一种把高斯基函数作为隶属函数的模糊神经网络视觉伺服控制器,该控制器不仅具有定性知识的表达能力,而且具有很强的在线学习能力。将所提方法用于三关节机器人"手-眼"视觉伺服系统的控制中,取得了很好的控制效果。

For solving the uncertainty problems in robotic visual servo system, a fuzzy neural network visual servo controller is presented. This controller combines fuzzy logic control and neural network together by using neural network to realize fuzzy inference, and using Guassian function as the membership function. The presented controller not only can express linguistic control knowledge, but also has good online learning ability, it is applied on a three-joint ＂eye-in-hand ＂ robotic visual servo system, and good control performance is achieved.