非结构化助老助残机械手对目标识别定位系统的研究

研制了一种非结构化的五自由度助老助残机械手,针对助老助残机械手末端手掌对空间目标物体的定位抓取问题,采用了一种激光引导的双目机器视觉定位和广义回归神经网络求逆解相结合的方法。该方法中,双目机器视觉定位采用双目工业相机,通过阈值分割算法识别激光斑确定目标物体的坐标,再经广义回归神经网络反解出助老助残机械手关节角度。最后通过采集实验数据和理想数据做对比,该方法能够达到较高的定位抓取精度,比较以往的机械手角度反解,该方法不存在无解或多解的情况。

Study on Unstructured Help Age Assistive Manipulator on the Target Recognition and Positioning System

A kind of manipulator with five freedom degree for the elderly and the disabled is developed witch work at unstructured state. In order to solve the problem of grasping the robot hand end palm to the space object, the method of combining the laser guided binocular machine vision location and the generalized regression neural network is used to solve the problem. In this method, binocular machine vision positioning using binocular industrial camera, through threshold segmentation recognition algorithm of laser spot to determine the coordinates of the object, the generalized regression neural network inverse solution for the elderly and the disabled mechanical hand joint angle. Finally, through the comparison of the experimental data and the ideal data, the method can achieve high positioning accuracy, compared with the previous mechanical hand angle inverse solution, the method does not exist no solution or multiple solutions.