| This dissertation deals with the kinematic calibration of a novel 3 DOF spindle head named A3 head, covering position analysis, error modeling, parameter identification, error compensation, computer simulation and experimental verification. The following creative work has been completed.On the basis of forward and inverse position analysis, the error model of 3-RPS mechanism is formulated by projecting the loop closure equations onto in the directions of actuation and constraints of each limb, allowing the geometric source errors affecting the compensatable and uncompensatable pose accuracy to be separated. Furthermore, the sensitivity analysis is carried out to provide a guideline in the manufacturing and assembly.The identification model using DBB is achieved, and then a two-step kinematic calibration algorithm is carried out, i.e. 1) to identify the relative and absolute home errors of each limb using a diameter and cylindrical gauge; 2) to identify the rest source errors using a DBB. The confidence degree of identified results can be enhanced in this algorism for the errors to be identified in each step are in the same order.The error compensation method is carried out by modifying the inputs of the system, and the assembly locations and number of DDB are determined to guarantee the robustness and identifiability. Computer simulation results show that the method proposed is robust.The effectiveness of the proposed approach is verified by experimental results. By using two-step kinematic calibration algorithm, the roundness error of the end point is reduced from 498μm to 192μm through home adjustment, and again reduced to 55μm by the further calibration using DBB. |