Research On Precise Positioning And Insertion Control Of Robot Hand-eye Cooperation System

Hand-eye cooperation based industrial robot is the key equipment of modern precision parts manufacturing and assembly industry.With the development of small batch,multi types and precision products,new requirements bring new challenges to robot precision positioning and insertion technology.Traditional hand-eye configuration modes and control algorithms are mainly based on the vision sensor installed at the end of the robot or independent of the manipulator,and few researches focus on the vision sensor installed in the middle joint of the robot.With the diversified development of robot types and application scenarios,the middle joint sensor configuration mode with high accuracy and low bias load has gradually become one of the important forms of cooperative system.It is accompanied by the low applicability and mismatch of traditional hand-eye calibration,visual servo and multi-sensor fusion algorithms.It is necessary to design new algorithms to realize the precise positioning and insertion of the hand-eye system under the new configuration.In this paper,a typical hand-eye system with vision sensor in the middle is studied.The main research contents are as follows:(1)Based on the configuration analysis of han-eye system,a reverse projection calibration method for hand-eye parameters is designed.Starting from the theoretical model,the configuration and calibration algorithm of hand-eye system are analyzed.The limitations of traditional calibration algorithm applied to the configuration of each joint axis parallel to each other are analyzed.Then,the rotation translation coupling characteristics of the system are described in detail,and the technical difficulties in the calibration process are clarified.The hand-eye system model and camera nonlinear geometric model based on geometric decomposition are established.On this basis,a hand-eye calibration algorithm based on reverse projection is proposed,and the key parameters affecting the system calibration accuracy are defined.Simulation and experiment are carried out to verify the feasibility and effectiveness of the proposed algorithm.(2)A simultaneous optimization algorithm for hand-eye calibration and point position measurement is proposed.Firstly,the effect of error transmission on measurement accuracy is analyzed.The degradation expansion model of the hand-eye system is established.Furthermore,a simultaneous optimization algorithm combining closed solution with nonlinear iterative optimization solution is proposed.The constraints between the number of samples and the measurement points are defined,and the solvability of the system model based on the number of conditions is analyzed.Finally,the error transfer experiments are carried out,and the calibration results and measurement errors are given,which verify that the proposed model and algorithm can achieve the simultaneous optimization of hand-eye parameters and feature coordinates.(3)Aiming at the problem of missing and inaccuracy of model parameters,a visual guidance control algorithm based on weak calibration model is proposed.A weak calibration algorithm for hand-eye parameters is proposed without the need of camera external parameters and calibration board.By analyzing the nonlinear and biaxial coupling characteristics of the system,the direction of error attenuation is obtained,and a weak decoupling visual guidance algorithm is designed.Lyapunov method is adopted to verify the stability of the algorithm.The constraints of control gain are defined.Finally,an adaptive gain adjustment algorithm based on the norm of pixel error is proposed.Comparative experiments show that the proposed algorithm can achieve visual tracking under the condition of missing model parameters.(4)In order to solve the problem that the degree of freedom of camera motion does not match the degree of freedom of robot,a robot precise positioning algorithm based on complementary visual servo is proposed.Firstly,the visual tracking constraints introduced by the mismatch of the number of degrees of freedom are analyzed.Based on the proposed motion constraints,the robot operation path is planned for model-free depth estimation.On this basis,a visual servo algorithm with independent control of each axis is proposed.The feedforward compensation of coupling term is added and the algorithm fusion in complementary space is carried out.Four-axis cooperative control experiments verify the feasibility and effectiveness of the proposed algorithm.(5)Aiming at a kind of typical electronic component insertion process,a compound insertion control algorithm based on visual and force information is proposed.Based on the analysis of the force characteristics of the components,the criteria for determining the contact state are designed,and the fallback adjustment strategy and mismatch alarm mechanism are proposed.The insertion state of the process of double displacement fusion is clarified.An interactive compensation control algorithm combining visual and force information is proposed.The effectiveness of the algorithm is verified by noise analysis,deflection experiment and force contact experiment.Finally,the electronic component insertion is completed by fusing multi-sensor information.Finally,the research contents of the paper are summarized and the future research is prospected.