Research On Sensorless Zero-force Control Of Collaborative Robots

"The 14 th Five-Year Plan proposes that by 2025,China will become a global robotics technology innovation curator,a high-end manufacturing gathering place and a new highland of integrated applications.And the collaborative robot field will usher in a period of vigorous development in the next few years.In order to reduce the difficulty of robot teaching and improve the efficiency of human-robot collaboration,dragging teaching has gradually become a research hotspot for collaborative robots.Zero-force control is an important technology to realize the dragging and teaching function,which can effectively reduce system complexity and save cost.At present,zeroforce control mainly relies on force/moment sensors.And without external force/moment sensors,the accuracy of the dynamics model will directly affect the compensation effect of zero-force control.Therefore,in this paper,zero-force control without force/moment sensors is investigated for collaborative robots.The details of the study are as follows:(1)A collaborative robot dynamics model was established.The robot dynamic model was developed based on the Newton-Euler iterative method by simplifying the structure of the BRTIRXZ0805 A 6-axis collaborative robot and deriving the robot dynamic model.To facilitate the realization of dynamic parameter identification,the linearized derivation of the kinetic model is carried out.Meanwhile,according to the model uncertainty analysis,nonlinear friction is introduced to improve the torque compensation accuracy.(2)A hybrid step-by-step nonlinear friction discrimination algorithm is given.The torque compensation accuracy of zero-force control depends on the discrimination accuracy of system dynamics parameters.A fifth-order Fourier series excitation trajectory is used to fully excite the robot dynamics,and the excitation trajectory coefficients are optimized by the genetic algorithm.A hybrid step-by-step nonlinear friction identification algorithm is proposed to improve the accuracy of joint friction torque calculation for the difficult problem of nonlinear friction model identification.This algorithm is based on the different sensitivity of different nonlinear friction coefficients to speed,and identifies the nonlinear friction model in steps to improve the accuracy of joint friction torque calculation.The effectiveness of the proposed algorithm is verified by simulation and experimental study with the robot hardware platform.(3)A zero-force control method for jointed sensorless robots is proposed.To improve the suppleness of the robot dragging demonstration,the robot dragging force is estimated based on the inverse dynamics model.And the zero-force control scheme combined with the moment model compensates for the robot’s gravity and friction moments.Based on this,the Coriolis moment and the nonlinear friction moment are compensated twice to reduce the required dragging moment and to improve the operator dragging physical sensation.(4)A zero-force coating system for collaborative robots was constructed.To verify the applicability of zero-force control of the articulated sensorless robot,an automated spraying control system was built based on vision servo control and zero-force control technology.The experimental results verified the feasibility and effectiveness of the self-correction method of the spraying robot trajectory attitude by estimating the current position of the spraying workpiece with a Mekamand camera and performing attitude correction of the spraying path teaching point.