Method Studies On Error Modeling And Calibration Of The Stewart Platform

The parallel robots are important to the advanced manufacturing. But there is a choke point in achieving parallel robots'industrialization, that they have a problem in kinematic precision. In this thesis, the research focuses on the error modeling and calibration of a Stewart platform.Based on analyzing the buildup of the platform, the main error source is studied and parameters in the error model are confirmed.The D-H coordinate systems are placed at every kinematic joint in each of the six chains, and then connections among these joints'coordinate are confirmed. Using the method of kinematic influence coefficient, the error influence coefficient is defined, and the kinematic error model is developed.The software of Matlab is used in error analysis programming, and error influence curves in some key points are given.The calibration modeling is developed, which is using the kinematic difference equations. Within the modeling the pose error of the joint points, the length error of the hydraulic cylinder and the error between the base coordinate and control coordinate are considered.Gauss-Newton method is used in the parameter identification programming. A estimating target for calibration is instituted. Kinematic pose data can be measured by DMM, then the error matrix is received which can be used to parameter identification. The result of the identification is used to error compensation.