A Two Step Calibration Method Of 6R Robot Based On Orientation And Distance Constraints

Industrial robots have been widely used in manufacturing field.The excellent property of high repeatability,high internal sensors resolution and complex trajectory movement facilitation enable them to fit special working conditions.However,their disadvantages are also obvious.The open-loop structure makes the geometric errors of each component joints,which are generated in manufacturing and assembling process,have a great impact on robot kinematic accuracy.It is difficult to meet the requirements of high orientation and position accuracy.Kinematic calibration is an effective method to improve the motion a ccuracy of robot.Current robot calibration research mainly focuses on position measurement,and the improving of motion accuracy after calibration is limited due to the relatively large identification errors of joint angle parameters.This paper proposes a6 R robot two-step kinematic calibration method which is intended to improve the kinematic parameters identification accuracy of current position constraint-based methods.In step 1,a dual-axis tilt sensor attached to the mounting plate of the robot is utilized to build orientation constraint,then identify and compensate angular parameters errors.In step 2,with the residual angular parameter errors be limited to a tiny range in solving process,a camera attached to the end-effector of the robot is utilized to build distance constraint and identify all identifiable kinematic parameters.In the calibration process,under the condition that all second order minute errors are omitted,the equation among joint parameter errors and robot end effector position errors is deduced based on the principle of robot differential motion;Based on the measurement principle of dual-axis tilt sensor,the pose error model is extended to the inclination error model and the relationship between joint angular parameter error and the robot end effector inclination is constructed.Proposed a convenient measure method to eliminate the non-parallelism between the xoy plane of robot base coordinate system and the horizontal datum of inclination sensor;Secondly,unidentified joint angular parameters are eliminated through QR decomposition;The reciprocal of the conditional number was taken as the observation index,20 poses were selected based on GA(Genetic Algorithm)and simulation study demonstrate that measurement poses optimal will minimize the impact of measurement errors.Thirdly,the distance constraint is constructed through camera,laser distance sensor and calibration plate.Then jo int parameter errors,including joint length parameter and residual angular parameters,are identified and compensated.Simulation study demonstrates that joint parameters identify accuracy of two-step calibration method is much higher than that of position-constraint-based method.After calibration,the motion accuracy is validated through dual-axis tilt sensor,gauge block and laser distance sensor.The average inclination error of the robot end effector x and y axis was reduced from 0.371° and 0.418° to 0.017° and 0.010°respectively,the average distance error of robot motion was reduced from 0.35 mm to 0.09 mm.