【摘要】 本文提出了机器人弧焊的路径修正策略。CCD摄像机用来获取实际焊缝图象 ,通过图象处理得到实际焊缝与预先示教路径之间的偏差。模糊神经网络用作推理系统。为了减小系统的稳态误差 ,考虑了偏差积分项的影响。三级插补方法用于实时路径修正。本文提出的策略得到实验结果的验证。更多还原

【Abstract】 A path correction scheme for robotic arc welding was presented. A CCD camera was used to obtain the image of actual weld seam. The deviation between actual weld seam and pre-taught path was derived in image processing. Fuzzy neural network worked as inference system. Integral term of deviation was also taken into account to reduce steady-state error. The three level interpolation method was adopted for real-time path correction. The validity of the scheme was verified by experiment results.更多还原