Study On Peg-in-hole Compliant Assembly Strategy And Method For Dual-arm Robot Based On Force Sensor

With the continuous improvement of industrial automation,industrial robots are widely used in peg-in-hole assembly tasks.In the process of relative motion between the peg and the hole,contact force and torque need to be controlled in a reasonable and acceptable range to complete the assembly smoothly.Thus,a compliant dual-arm robot is introduced,and a six-axis force sensor is to detect the force and torque directly.Force information is used to guide the coordinated motion of dual-arm of the robot.The main contents of this paper are as follows:First,forward and inverse kinematics models of the dual-arm robot are analyzed,and the working space of dual-arm is constructed.Relationships of geometric and force in various contact conditions are analyzed,and the coordinated motion of dual-arm is obtained to ensure the peg-in-hole assembly smoothly.Secondly,on the basis of the two-phase assembly method,three kinds of reliable searching strategies based on force sensor are proposed to achieve the corresponding assembly tasks.The first is to assemble the round and square pieces with small angle deviations under random search hole.The method searches for the hole in a small range of contact planes in the case of force sense only.The second is the assembly of different shapes pieces with a larger angle deviation under the guidance of visual/force sense.This method takes vision sensor for rough positioning and force sensor for precise adjustment.The third is to assemble the different shaped pieces with arbitrary angle deviation by using the master and slave arm to coordinate the motion at any grasping positions.This method plans the motion of master arm using the current pose information of slave arm,which can overcome the influence of inaccuracy of the robot.Then,we carried out corresponding experiments on different assembly tasks to verify the effectiveness of the three kinds of assembly strategies.In addition,we also set up a comparative experiment to compare the assembly success rate and assembly efficiency of the single arm random searching,dual-arm random searching,vision and force coordination guided searching,and master and slave coordination search.The experimental results show that the single arm random searching has a certain blindness,the assembly time is the longest,and the force and torque control effect is the worst;the assembly time of dual-arm random searching is long;visual and force coordination guidance and the coordinated motion of master and slave arm have obvious advantages in the assembly efficiency and the assembly success rate.