Because of the welding complication of arc welding robot,it not only requires welding-gun at the right position, but also requires the welding-gun moving through the welding seam trace,so as to adapt to some restricting condition in geometry.And the choice of welding gesture is the key to get better welding quality.The extraction of welding-gun pose's information for arc welding robot is limited in right or left deviation and the height of welding-gun.The added sensor is used to extract the welding-torch's angle.This method is only solve the limit of geometry restrict,the angle infection to the welding quality and extraction of welding-gun pose's information from arc welding sensor information.So we do the research as follows:Firstly,we introduce the arc welding sensor rotated in high speed and the means to extract welding-torch gesture's information in and abroad.Then,simulating the welding-gun pose by UG software,getting the welding blow lamp height waveform image and finding every gesture's feature from the image provide theory foundation to real welding experiment.We do research in the welding experiments' interference,so as to compare every filter algorithm with each other.Finally,extracting of welding-torch pose's information by characteristic harmonic,finding out the rule provide the important information to auto control the welding-gun pose. |