Poses Selection Using Genetic Algorithm to Improve the Local Poe Kinematics Calibration

This paper investigates the use of genetic algorithm to optimize poses selection to improve kinematic calibration for manipulator. Genetic algorithm is used to determine the optimal poses while iterative least square algorithm is used to calibrate the kinematics model of the manipulator. Observability index are used to evaluate the optimality of the set of poses. The fitness function of genetic algorithm is chosen from the observability index. In addition, local POE (Product of Exponential) method is used to model the manipulator kinematics. The objective of this paper is to design an algorithm which optimizes the number of poses while improving the calibration performance. The experiments utilize 7-DOF Mitsubishi PA-10 manipulator as the platform and a LEICA laser tracker as the measurement tool. The experiment shows that genetic algorithm can optimize the number of poses and improve the calibration performance Keyword: Genetic Algorithm, Kinematics Calibration, Local POE, Pose Optimization Introduction Precise position control requires an accurate model of the manipulator. However, in practice, the accuracy of the kinematics model is reduced due to several external errors such as: manufacturing errors, link misalignment, and assembly errors at the manipulator. Kinematics calibration is presented as a solution to improve the accuracy of the manipulator. In general, kinematic calibration is influenced by the calibration algorithm, the modeling method, the quality of the poses and the measurement device. There are several methods for kinematics calibration in literature. Wang, C,B et al [1] made use of a forward calibration method. The forward calibration identifies the actual parameter of the manipulator based on the measurement at the workspace of the manipulator. Even though the method shows promising improvement, it becomes a problem if the inverse model of the manipulator is needed. Doria, A et al [2] introduced inverse kinematics using B-splines and multivariate parametric approximating splines functions as tools to do calibration. Chen, I.M, et al [3] proposed a least square method to calibrate the manipulator. Local product of exponential is used to model the kinematics. Several works related to this method are also proposed in [4],[5], and [6]. Although the local POE calibration method provides significant improvement, this technique requires a lot of poses. To optimize the poses for calibration, several techniques are introduced. Chernoff, H [7] minimized the trace of non-zero singular values of the Jacobian matrices of the manipulator. This method is called A-Optimality. Wald [8] maximized the determinant of the non-zero singular values of the Jacobian matrices of the manipulator. This method is introduced as D-Optimality. Smith [9] proposed G-Optimality, it minimizes the maximum prediction variance of the non-zero singular values of the Jacobian matrices. Ehrenfeld [10]