This paper focuses on the kinematic design issues for a three degrees-of-freedom (DOFs), i.e., two-rotational and one-translational (2R1T) end-effector to perform continuous contact operations, such as deburring, grinding, and polishing. The proposed end-effector design employs a 3-legged prismatic-prismatic-spherical (3-PPS) parallel mechanism due to its desired kinematic characteristics and dynamic behavior. As the 3-PPS parallel mechanism is featured with zero-torsion motion characteristic, the orientation of its moving platform can be always represented by a rotation about an axis parallel to its base platform plane. Through analysis of the rotation matrix of the moving platform, closed-form linear solutions for both forward and inverse displacement analyses are readily derived. Other critical design issues, such as passive prismatic joint displacement, parasitic motion, velocity, and singularity analyses, are addressed. For a specific dimension design of the 3-PPS parallel mechanism, the workspace analysis indicates that the proposed design can achieve a singularity-free ±12°×±12°×25 mm workspace. Furthermore, as the displacements of the passive prismatic joints are within 2.63 mm, light-weight flexure-based prismatic joints are designed to replace the conventional heavy linear guides. The flexure-rigid structure of the 3-DOF 2R1T end-effector significantly improves the dynamic performance of the system. A prototype of the 3-DOF 2R1T robotic end-effector is designed and fabricated to verify the proposed design.

High-quality and diverse datasets are crucial for supervised learning in industrial defect detection, yet collecting such data remains challenging. Synthetic data generation offers a promising solution for data augmentation, but traditional methods struggle with controlling key semantic features, leading to domain gaps that impair downstream performance. To address these challenges, we introduce DD-Aug, a novel data augmentation pipeline that integrates Cinema 4-D (C4D) and stable diffusion (SD) to generate photorealistic synthetic images. C4D is used to model key semantic features, such as contours, textures, and lighting conditions, while SD performs style transfer to increase the realism of generated images and reduce the domain gap. DD-Aug translates human visual knowledge into synthetic images, offering precise control over semantic features while adhering to physical rules and reducing the need for complex prompt engineering. This approach enhances generalization, making it particularly suitable for challenging industrial imaging conditions. By balancing semantic consistency and data diversity, DD-Aug significantly improves the quality of synthetic data. Our evaluations on the deformable defect bearing dataset (NBBD) demonstrate that DD-Aug surpasses other generative augmentation methods, yielding a 2.5%–10% improvement in detection accuracy across multiple advanced object detectors compared to models without augmentation. These results demonstrate DD-Aug's potential to significantly advance industrial defect detection through improved synthetic data quality and performance.

Free-piston linear generators (FPLGs) have attractive application prospects for hybrid electric vehicles (HEVs) owing to their high-efficiency, low-emissions and multi-fuel flexibility. In order to achieve long-term stable operation, the hybrid system design and full-cycle operation strategy are essential factors that should be considered. A 25 kW FPLG consisting of an internal combustion engine (ICE), a linear electric machine (LEM) and a gas spring (GS) is designed. To improve the power density and generating efficiency, the LEM is assembled with two modular flat-type double-sided PM LEM units, which sandwich a common moving-magnet plate supported by a middle keel beam and bilateral slide guide rails to enhance the stiffness of the moving plate. For the convenience of operation processes analysis, the coupling hybrid system is modeled mathematically and a full cycle simulation model is established. Top-level systemic control strategies including the starting, stable operating, fault recovering and stopping strategies are analyzed and discussed. The analysis results validate that the system can run stably and robustly with the proposed full cycle operation strategy. The effective electric output power can reach 26.36 kW with an overall system efficiency of 36.32%.

Removing the parameter redundancy in the kinematic error model does improve the robustness of parameter identification. However, this does not help much on saving the calculation time on the error compensation process, which is closely related to the data structure used for representing rigid-body motion. Compared with the homogenous transformation matrix (HTM), the unit dual quaternion (UDQ) can describe the rigid-body motion in term of an array with eight entries, thus it is a conducive data structure for efficient kinematic computation. In this article, the exponential and logarithm mappings of an UDQ are defined explicitly so that the relationship between finite and instantaneous motions is set up under the exponential coordinate. Thereafter, by defining the adjoint operator of UDQ to transform the twist under different frames, the linearized kinematic error models are built with the local product-of-exponentials (POE) formula. This facilitates the upcoming parameter identification and error compensation processes. Especially, the counts of elementary operations are evaluated throughout the compensation processes, which shows the number of arithmetic operations are greatly reduced compared with using the POE formula based on the HTM. Additionally, the simple data structure in term of 1-D arrays saves the data dispatch time when performing adjoint operations. These make the developed model suitable for calibration of the articulated robot with a long sequence of trajectories. Experiments on a 6-degree-of-freedom industrial robot validate the effectiveness of the proposed approach.

Installing force-controlled end-effectors on the end of industrial robots has become the mainstream method for robot force control. Additionally, during the polishing process, contact force stability has an important impact on polishing quality. However, due to the difference between the robot structure and the force-controlled end-effector, in the polishing operation, direct force control will have impact during the transition from noncontact to contact between the tool and the workpiece. Although impedance control can solve this problem, industrial robots still produce vibrations with high inertia and low stiffness. Therefore, this research proposes an impedance matching control strategy based on traditional direct force control and impedance control methods to improve this problem. This method’s primary purpose is to avoid force vibration in the contact phase and maintain force–tracking performance during the dynamic tracking phase. Simulation and experimental results show that this method can smoothly track the contact force and reduce vibration compared with traditional force control and impedance control.