Structure Optimization And Parameters Calibration Of Cable-driven Parallel Robots

Cable-driven parallel robots(CDPRs)control the mobile platform by changing the cable length or the cable tension in the workspace.CDPRs have the characteristics of high speed,high load,and low motion inertia.Influenced by the flexible cables dis-persed in the workspace,there are still many challenges in the optimization design and parameter calibration of CDPRs.Therefore,this dissertation focuses on the geometry configuration selection,parameter optimization,parameter calibration and parameter sensitivity analysis of the CDPR.The main research of this dissertation include the fol-lowings:1)The working state of CDPR for automatic warehouse systems is analyzed.Con-sidering the wrench-feasible condition of the mobile platform,the maximum load crite-rion is designed which is based on the external wrench applied on the mobile platform-In order to further optimizing the energy consumption of the CDPR,the cable tension criterion is designed considering the wrench-feasible workspace(WFW).The constraint conditions of the geometry configuration are defined by using the constraint conditions of cable-cable,cable-platform and cable-structure collision and considering the geom-etry symmetry of the CDPR,which further reduce the search space of geometry con-figurations.The feasible geometry configuration and mobile platform parameters are optimized by using a stochastic optimization algorithm,and the optimization results are tested by simulations,and the distribution of optimization criteria in the workspace is obtained.The optimization of the structural configuration and the mobile platform pa-rameters are carried out by using the Tabu optimization algorithm,and the simulation experiments verified the optimized results.The maximum load criterion of the opti-mized mobile platform is increased by 37.8%,and the workspace and cable tension criteria were significantly improved.Finally,static equilibrium experiments were con-ducted on the laboratory-developed cable-driven parallel robot to verify that the mobile platform with the optimized selected configuration and parameters can significantly re-duce the average cable tensions.2)In order to improve the accuracy and robustness of parameters calibration,a measurement optimization selection algorithm is designed based on the tabu search al-gorithm.The mathematical properties of the five observability index are analyzed,and the best observability index is used to select the measurement poses.Then,based on the extended Kalman filter algorithm,a parameter calibration experiment was verified on the 6-DOF cable-driven parallel robot.Based on the randomly selected measurement poses,the average position error of the mobile platform is reduced from 48.37(mm)to 5.49(mm).Based on the optimized measurement pose,the mean position error of the mobile platform is reduced to 1.53(mm),and the variance is reduced to 0.24.In or-der to further improve the position accuracy of the cable-driven parallel robot,aiming at the measurement and calibration of the transformation matrix parameters between coordinate systems,a method for iterative identification of the transformation matrix parameters and geometric parameters is proposed,and the convergence of the iterative identification algorithm is proved.Further,considering that the residual position error of the cable-driven parallel robot is affected by the non-geometric parameter error,the artificial neural network is used to compensate for the residual position error caused by the non-geometric parameter error.In order to improve the generalization ability and fitting accuracy of the neural network,a neural network optimization method based on a hierarchical genetic algorithm is proposed to optimize the network structure and parameters.The non-geometric parameter errors are compensated in the joint space of the robot.Finally,a parameter calibration experiment was carried out on 3-DOF cable-driven parallel robot.The geometric parameters are calibrated using the proposed it-erative identification method.The maximum error of the mobile platform position is reduced from 61.5(mm)to 10(mm).Through non-geometric parameter error compen-sation,the maximum error of the mobile platform during the movement is reduced from 13.4(mm)to 3.2(mm).3)In order to improve the obstacle avoidance ability and adaptability of CDPRs,a new configuration continuous RCDPR is proposed.Traditional constraints on the mo-bile platform are replaced by pulley-based constraints,and the geometric parameters are adjusted to improve the load capacity and reconfigurability of the RCDPR.According to the screw theory,the RCDPR is modeled and analyzed.Furthermore,the solution method of the inverse kinematics of the RCDPR is given,and a multi-solution method for solving the inverse kinematics is designed based on the interval analysis method.The optimization criterion of the off-line solution considering the stiffness and the ca-ble tension is designed,and the stochastic optimization algorithm is used to optimize the continuous reconfigurable position.Using the energy consumption of the RCDPR as the optimization criterion of the online solution,the dynamic performance of the mobile platform is used to constrain the solution space and discrete step of the reconfigurable position,so that the online solution speed can meet the real-time motion planning and obstacle avoidance requirements.Through the singularity analysis of the Jacobian ma-trix,the singular conditions of the RCDPR are obtained.The cable tension factor in the WFW is proposed as the optimization criterion,and the particle swarm optimization algorithm is used to optimize the mobile platform parameters of the RCDPR.4)In order to improve the model accuracy of the RCDPR,this dissertation models the position of the take-out point of the cable on the pulley.By analyzing the first-order differential model and the second-order differential model of the cable and the mobile platform,the velocity and the acceleration mapping from the mobile platform to the cables is obtained.The continuous speed and acceleration of the cable verified the effectiveness of the proposed RCDPR.The reliability theory is proposed to analyze the sensitivity of geometric parameters and the parameters reliability of cable length.The response surface method is used to model and analyze the parameters reliability of cable length,and the solution methods of parameters reliability and parameter sensitivity are given.By analyzing the sensitivity of geometric parameters,it provides a theoretical basis for parameter optimization and parameter calibration of RCDPRs and improves the efficiency of optimization design and parameter identification for RCDPRs.