Recently, robot arms have become an irreplaceable production tool, which play an important role in the industrial production. It is necessary to ensure the absolute positioning accuracy of the robot to realize automatic production. Due to the influence of machining tolerance, assembly tolerance, the robot positioning accuracy is poor. Therefore, in order to enable the precise operation of the robot, it is necessary to calibrate the robotic kinematic parameters. The least square method and Levenberg-Marquardt (LM) algorithm are commonly used to identify the positioning error of robot. However, it generally has the overfitting caused by improper regularization schemes. To solve this problem, this article discusses six regularization schemes based on its error models, i.e., $L_{1}$ , $L_{2}$ , dropout, elastic, log, and swish. Moreover, this article proposes a scheme with six regularization to obtain a reliable ensemble, which can effectively avoid overfitting. The positioning accuracy of the robot is improved significantly after calibration by enough experiments, which verifies the feasibility of the proposed method.

With the rapid development and wide applications of industrial manipulators, a vital concern rises regarding a manipulator's absolute positioning accuracy. The manipulator calibration models have proven to be highly efficient in improving the absolute positioning accuracy of an industrial manipulator. However, existing calibration models commonly suffer from the low calibration accuracy caused by the ignorance of non-geometric errors. To address this critical issue, this paper proposes an E xtended Kalman Filter-incorporated R esidual Neural Network-based C alibration (ERC) model for kinematic calibration. Its main ideas are two-fold: a) adopting an e xtended Kalman filter to address a manipulator's geometric errors; and b) adopting a r esidual neural network to cascade with the e xtended Kalman filter for eliminating the remaining non-geometric errors. Detailed experiments on three real datasets collected from industrial manipulators demonstrate that the proposed ERC model has achieved significant calibration accuracy gain over several state-of-the-art models.

Industrial robot arms are the critical equipment of industrial production and have been widely adopted in various fields, such as the aerospace and petrochemical industry. However, an uncalibrated robot arm suffers from extremely low absolute positioning accuracy, which cannot satisfy the accuracy requirements of high-precision manufacture. To address this thorny issue, it is extremely important to implement periodic calibration for robot arms. In general, most experts on calibration systems have rich mechanical and instrument experience. However, due to the particular complexity involved in the collection of calibration data, it is exceedingly difficult for experts in other fields to conduct robot arm calibration studies. In this work, we consider the issue of robot arm calibration from the perspective of machine learning and develop a publicly available dataset called "RobotCali". The main ideas of this work are four-fold: a) developing a publicly available dataset to assist researchers from other fields in conducting calibration experiments and validating their ideas, b) adopting an extended Kalman filter algorithm to suppress the measurement noises in a robot arm calibration system, c) designing an improved covariance matrix adaptive evolution strategy to achieve fast convergence rate and high searching stability, d) proposing a novel calibration system based on an extended Kalman filter and an improved covariance matrix adaptive evolution strategy. Additionally, extensive experiments demonstrate that compared with state-of-the-art calibration systems, the proposed calibration system obtains a highly competitive calibration accuracy.

Robots facilitate a critical category of equipments to implement intelligent production. However, due to extensively inevitable factors like structural errors and gear tolerances, the positioning error of an industrial robot is several millimeters, therefore failing in fulfilling the high-precision manufacture requirements. To address the critical problem, this work develops a novel calibration algorithm incorporates an Unscented Kalman Filter and a Variable Step-size Levenberg-Marquardt (UKF-VSLM) algorithm for efficient industrial robot calibration with the following two-fold ideas: a) developing a novel Variable Step-size Levenberg-Marquardt algorithm to address the local optimum issues encountered by a standard Levenberg-Marquardt algorithm; and b) incorporating an unscented Kalman filter into the proposed Variable Step-size Levenberg-Marquardt algorithm to suppressing the measurement noises during the calibration process. Empirical studies on an HSR JR680 industrial robot demonstrate that compared with state-of-the-art calibration algorithms, the calibration accuracy of the developed UKF-VSLM is 19.51% higher than that of the most accurate Levenberg-Marquardt algorithm measured by the maximum error. The empirical results strongly support the superior performance of the proposed algorithm in addressing robot calibration issues.

During the past decades, industrial robot arms have been widely adopted in various fields, like aerospace and medical care. However, an uncalibrated industrial robot arm suffers from its high absolute positioning error, which tremendously restricts its application in high-precision intelligent manufacture. Related robot arm calibration methods present a potential solution to this issue, while recent algorithms frequently encounter some issues, such as long-tail convergence and low calibration accuracy. To address this thorny issue, this study proposes a novel robot arm calibration method incorporating the extended Kalman filter with Quadratic Interpolated Beetle Antennae Search algorithm. It adopts three-fold ideas: a) developing a new Quadratic Interpolated Beetle Antennae Search algorithm that greatly addresses the issue of local optimum and low convergence rate in a Beetle Antennae Search algorithm; b) adopting an extended Kalman filter algorithm to decrease the influence of measurement noises and c) proposing a robot arm calibration method incorporating an extended Kalman filter with Quadratic Interpolated Beetle Antennae Search algorithm to searching the optimal kinematic parameters of a robot arm. Extensive experiments on an ABB IRB120 industrial robot arm demonstrate that the calibration accuracy of the proposed method is 5.17% higher than that of the state-of-the-art method. Therefore, it is appropriate for practical application occasions with high-precision requirements.

The absolute positioning accuracy of an industrial robot arm is vital for advancing manufacturing-related applications like automatic assembly, which can be improved via the data-driven approaches to robot arm calibration. Existing data-driven calibrators have illustrated their efficiency in addressing the issue of robot arm calibration. However, they mostly are single learning models that can be easily affected by the insufficient representation of the solution space, therefore, suffering from the calibration accuracy loss. To address this issue, this study proposes a calibrator fuzzy ensemble (CFE) with twofold ideas: 1) implementing eight data-driven calibrators relying on different sophisticated machine learning algorithms for an industrial robot arm, which guarantees the accuracy of individual base models and 2) innovatively developing a fuzzy ensemble of the obtained eight diversified calibrators to obtain impressively high calibration accuracy for an industrial robot arm. Extensive experiments on an ABB IRB120 industrial robot implemented with MATLAB demonstrate that compared with state-of-the-art calibrators, CFE decreases the maximum error at 8.59%. Hence, it has great potential for real applications.