Advanced Robotic Kinematic Parameter Calibration Based On Artificial Bee Colony Algorithm

Nowadays,with the advent of the era of industrial 4.0,firms are more inclined to the combination of industrialization and intelligent,and more robots are put into various fields of industrial production.At the same time,the requirement of the robot accuracy also higher than ever.It has become an urgent problem in the field of robotics to improve the robot's accuracy by calibrating kinematic parameters of the robot.In previous studies of calibration on kinematic parameters,only few paper have considered the influence of measurement disturbance on accuracy of robot kinematics calibration.In this paper,in order to improve calibration's anti-jamming of measurement disturbance,the existing method of the robot kinematics calibration is improved and a new method of robot kinematics calibration based on artificial bee colony(ABC)algorithm is put forward.There are two improvements: firstly,in the experiments,the joint angles which determine the end effector's orientation and position are optimized,which improves the anti-jamming of measurement disturbance;secondly,a new method of hand-eye calibration is proposed,which improves the accuracy of hand-eye calibration.An experiment based on artificial colony algorithm is designed,and a Kinect camera is used as the main measuring tool.Details are described as follows:Firstly,the forward kinematic model using nominal DH parameters of the robot was established,and the orientation and position of the end effector was then obtained.On this basis,kinematic error model of this robot was then established,that is,the mathematic and linear relationship between position error of the end effector and the joint parameter error.A group of measurement joint angles which increases the anti-jamming performance of the calibration experiment was obtained using Artificial Bee Colony algorithm,that is,a new kind of intelligence algorithm which is widely used in system identification.Then the camera which is used in calibration experiment was calibrated.In the camera calibration experiment,Kinect camera moved along with the end effector and took some pictures of the calibration chess board which located in the range of visibility of the camera.The inner parameters of Kinect were obtained using relative mathematic image algorithm.As a side-effect of camera calibration,the relative position and orientation between camera and chess board also can be obtained.Next,hand-eye calibration which is used to get the relative position and orientation between camera and the end effector was carried out.In order to solve the problem that the hand-eye calibration result cannot be directly used in kinematic parameter identification,an advanced method of hand-eye calibration based on ABC algorithm was proposed.A hand-eye matrix with more accuracy was obtained,and can be directly used in kinematic identification without normalization.Together with the relative orientation and position between camera coordinate and world coordinate obtained in the previous step,the one between the coordinate of the end effector and that of the world can also be obtained.Finally,together with robotic error model established in the first step,relative orientation and position between the end effector and the world obtained from the previous step and the nominal DH parameters of the robot which is needed to be calibrated,kinematic parameters was obtained using Least Square method.In order to verify the reliability of the parameters,a confirmatory experiment was designed and carried out.The result of the experiment showed that the advanced calibration method proposed in this paper is reliable and feasible.