Design of experiments for calibration of planar anthropomorphic manipulators

Abstract

-The paper presents a novel technique for the design of optimal calibration experiments for a planar anthropomorphic manipulator with n degrees of freedom. Proposed approach for selection of manipulator configurations allows essentially improving calibration accuracy and reducing parameter identification errors. The results are illustrated by application examples that deal with typical anthropomorphic manipulators.

Optimization of Measurement Configurations for Geometrical Calibration of Industrial Robot

“…It can be shown that here it is not reasonable to use the A-criterion (the goal of A-criterion is to minimize the trace of the covariance matrix) because the trace of the relevant information matrix does not depend on the plan of experiments. Further, it was proved that the criteria that operate with the covariance matrix determinant (D and D* criteria, [8]; the goal of D-criterion is to minimize the determinant of the covariance matrix, the goal of D*-criterion is to ensure independence of the identified parameters and to minimize the determinant of the covariance matrix that is diagonal) lead to minimization of only. Corresponding loss of the accuracy is presented in Table 1.…”

Geometrical Patterns for Measurement Pose Selection in Calibration of Serial Manipulators

“…of the covariance matrix or information matrix (its inverse) [1]. In [5] the authors proposed an optimality condition of the calibration plan where the main idea is to ensure that the information matrix is diagonal. This condition corresponds to the D-optimal plan of experiments and is satisfied when…”

“…In fact, for n = 2, m ≥ 2 expressions (4) provide two simple constraints only ∑ m k=1 cos q k 2 = 0 , ∑ m k=1 sin q k 2 = 0, which are imposed on the 2m design variables q k 1 , q k 2 | k = 1, m . So, relevant configurations can be easily found geometrically [5]. However, even in the case of n = 3, m ≥ 3 there are 6 rather complicated trigonometric constraints imposed on 3m design variables q k 1 , q k 2 , q k 3 | k = 1, m .…”

Identification of the manipulator stiffness model parameters in industrial environment

“…Daney [53] used the constrained optimization algorithm based on the minimization of the singular values product for Gough platform calibration. In [54], the idea from D-optimality criteria has been used to quantify optimal measurement configurations for a general nR planar manipulators. Zhuang et.…”