Measurement Error Analysis And Compensation Of Passive Space Position Tracking Measurement System

Laser tracker is the main instrument for calibrating the robot’s spatial positioning accuracy.However,the structure of the laser tracker is complex and expensive,and it is difficult to popularize it in the production of enterprises.Therefore,it is very necessary and urgent to develop new structures or methods to reduce manufacturing costs and popularize them in industrial measurement.The research group has developed a passive space position tracking measurement system with simple structure and low cost,which is used to realize the space position measurement of industrial robots.In order to improve the measurement accuracy of the measurement system,this paper analyzes the geometric errors of the measurement system,proposes errors measurement methods,establishes errors compensation models,and proves the effectiveness of the errors compensation models through verification experiments.The research contents of this paper mainly include:(1)The geometric error of passive space position tracking measurement system is analyzed,and an error model based on screw theory is established.The geometric error terms of the measurement system are described by position-dependent and position-independent errors.Rigid motion can be described as a motion screw,and the transformation matrix of rigid motion is obtained through the exponential transformation of the motion screw.Based on the above screw theory,the ideal kinematic model and geometric errors compensation model of the passive space position tracking measurement system are established.(2)Design experiments to measure and calibrate the error parameters.The angle measurement errors of the circular grating embedded in the turntable is calibrated using a polyhedral prism,and the angle measurement error is compensated by using the harmonic function to fit discrete values.The position-independent errors between the vertical rotation axis and the horizontal rotation axis is measured using a multimeter.The positionindependent errors between the telescopic axis and the horizontal rotation axis is calibrated by using a combination of laser collimation and positive-inverse mirrors.(3)Design experiments to verify the validity of error modeling based on screw theory.Use a coordinate measuring machine(CMM)as a standard comparison instrument.In the measurement space of 450×450×200mm,the coordinate data of 180 measurement points are collected.A linear model is used to solve the transformation matrix to realize the transformation from the coordinate system of the measuring system to the coordinate system of the coordinate measuring machine.Finally,the measurement error between the measured value at the measurement point and the theoretical value is calculated.After compensation based on the error model of screw theory,the maximum measurement error is reduced from611.9 μm to 132.2 μm,and the compensation rate is 78.4%.(4)The BP neural network error model optimized by genetic algorithm is established and the verification experiment is carried out.After compensation of the error model based on screw theory,the measurement error of the measurement system is still relatively large.The BP neural network model optimized by genetic algorithm is used to predict and compensate the residual error.A BP neural network model is established,and the parameters in the model are optimized by genetic algorithm.Taking the spherical coordinate parameters of the passive space position tracking measurement system as the input and the residual error as the output,the BP neural network model optimized by the genetic algorithm is trained.After the training is completed,the BP neural network prediction model optimized by the genetic algorithm is successfully established.Then,after compensation by the BP neural network model optimized by genetic algorithm,the maximum measurement error is reduced from 132.2 μm to 57.7 μm,and the compensation rate is 56.4%.