Visual Positioning Method For Industrial Robot Peg-in-Hole Assembly

As the important automation equipment of modern manufacturing,industrial robot has been widely used in conveying,spraying,welding,assembly,and other production tasks.With the requirement of robot's intelligence and autonomy in modern manufacturing,intelligent application of industrial robot has become a research focus in current robot technology.Component assembly is a common and important task in production process.And researching a kind of target recognition and positioning method which used in peg-in-hole assembly process of flexible production unit has great significance in improving intelligent application level of industrial robot and production efficiency.This paper takes a six axis serial industrial robot as the research object,setting up a experiment platform for target recognition and visual positioning.And the problems of visual positioning and automatic control in the assembly process with industrial robot are deeply discussed.For the six axis serial industrial robot,kinematic analysis and singular avoidance trajectory planning are researched in detail.Firstly,the kinematics model of SR10 C industrial robot is established using D-H(Denavit-Hartenberg)method,and the forward and inverse kinematics,jacobian matrix of the robot has been thoroughly analyzed.Then the correctness of the proposed closed-form solution has been verified through the simulation experiment.On this basis,the reasons which causing kinematics singularity of six axis serial industrial robot have been researched.The singular parameters of SR10 C industrial robot are separated and extracted and differential kinematics formulas with damping coefficients are deduced.Furthermore,the simulation experiment of singular avoidance trajectory planning are carried out.It avoids the situation appeared which movement position inaccessible because of the robot kinematics singularity.It proposed an visual positioning method for industrial robot peg-in-hole assembly.Through designing image processing and character recognition algorithm and servo control method of industrial robot,the whole visual positioning process are completed.First of all,using depth image achieves coarse positioning to quickly identify targets,then completing secondary positioning to ensure accuracy based on color image.This method can not only avoids the interference of image processing result which caused by target's appearance characteristics such as color and texture,but also don't be affected by the outside light and complex environment background.On this basis,a PC software which integrating image processing and robot control functions was developed.It establishes the connection between robot vision system and control system,and directly generates control commands using the data that obtained by vision system to realize the automation of the whole positioning process and monitor the running status of the system in real time.On the basis of good results of kinematics singularity avoidance simulation and target features extraction experiments,the visual positioning experiment for industrial robot peg-in-hole assembly was carried out.The experimental results show that this visual positioning method solved the target recognition and positioning problem in peg-in-hole automatic assembly process of flexible production unit,and avoided the environment noise interference to visual system.In the meantime,it has stable positioning accuracy and satisfies the requirement of the positioning error of follow-up peg-in-hole assembly process.And it has better flexibility and adaptability in the practical applications.