When the robot is positioned to grab workpiece,the correct of choosing workpiece feature parameters is the key to whether the robot can obtain the grabbing point of workpiece accurately and grab the workpiece. Based on the method of image processing, neural network nonlinear processing ability was proposed to solve the nonlinear problems in workpiece feature selection and feature extration. In the training of the neurons, by using the improved Hebb learning rule, the problem of unrestricted growth of weights and no convergence in traditional learning mode was overcome.The recognition degree of the feature and the accuracy of feature extraction are improved ,which makes the robot realize accurate scraping of the workpiece.
参考文献
相似文献
引证文献
引用本文
李振雨,王好臣,丁兆勇,王功亮.基于神经网络算法的机器人定位技术研究[J].机床与液压,2018,46(15):84-87. . Research on Localization Technology of Robot Based on Algorithm of Neural Network[J]. Machine Tool & Hydraulics,2018,46(15):84-87