• First beam image dataset capturing diverse optical-alignment patterns and parameters • Optical resonator alignment cast as a pairwise beam-pattern regression task • Two-stage model with both feature interaction and refinement for coarse-to-fine alignment • Trained on one device, the model generalizes to another without re-training • Achieves high accuracy with real-time inference on embedded edge hardware Accurate alignment of a laser resonator is essential for upscaling industrial laser manufacturing and precision processing. However, traditional manual or semi-automatic methods depend heavily on operator expertise, and struggle with the interdependence among multiple alignment parameters. To tackle this, we introduce the first real-world image dataset for automatic laser resonator alignment, collected on a laboratory-built resonator setup. It comprises over 6,000 beam profiler images annotated with four key alignment parameters (intracavity iris aperture diameter, output coupler pitch and yaw actuator displacements, and axial position of the output coupler), with over 500,000 paired samples for data‐driven alignment. Given a pair of beam profiler images exhibiting distinct beam patterns under different configurations, the system predicts the control-parameter changes required to realign the resonator. Leveraging this dataset, we propose a novel two-stage deep learning framework for automatic resonator alignment. In Stage 1, a multi-scale CNN augmented with cross-attention and correlation-difference modules, extracts features and outputs an initial coarse prediction of alignment parameters. In Stage 2, a feature-difference map is computed by subtracting the paired feature representations and fed into an iterative refinement module to correct residual misalignments. The final prediction combines coarse and refined estimates, integrating global context with fine-grained corrections for accurate inference. Experiments on our dataset and a different instance of the same physical system from which the CNN was trained suggest superior accuracy and practicality to manual alignment.

This paper studies 3D LiDAR mapping with a focus on developing an updatable and localizable map representation that enables continuity, compactness and consistency in 3D maps. Traditional LiDAR Simultaneous Localization and Mapping (SLAM) systems often rely on 3D point cloud maps, which typically require extensive storage to preserve structural details in large-scale environments. In this paper, we propose a novel paradigm for LiDAR SLAM by leveraging the Continuous and Ultra-compact Representation of LiDAR (CURL) introduced in [1]. Our proposed LiDAR mapping approach, CURL-SLAM, produces compact 3D maps capable of continuous reconstruction at variable densities using CURL's spherical harmonics implicit encoding, and achieves global map consistency after loop closure. Unlike popular Iterative Closest Point (ICP)-based LiDAR odometry techniques, CURL-SLAM formulates LiDAR pose estimation as a unique optimization problem tailored for CURL and extends it to local Bundle Adjustment (BA), enabling simultaneous pose refinement and map correction. Experimental results demonstrate that CURL-SLAM achieves state-of-the-art 3D mapping quality and competitive LiDAR trajectory accuracy, delivering sensor-rate real-time performance (10 Hz) on a CPU. We will release the CURL-SLAM implementation to the community.

Maps of LiDAR Simultaneous Localisation and Mapping (SLAM) are often represented as point clouds. They usually take up a huge amount of storage space for large-scale environments, otherwise much structural detail may not be kept. In this paper, a novel paradigm of LiDAR mapping and odometry is designed by leveraging the Continuous and Ultra-compact Representation of LiDAR (CURL) proposed in [1]. Termed CURL-MAP (Mapping and Positioning), the proposed approach can not only reconstruct 3D maps with a continuously varying density but also efficiently reduce map storage space by using CURL’s spherical harmonics implicit encoding. Different from the popular Iterative Closest Point (ICP) based LiDAR odometry techniques, CURL-MAP formulates LiDAR pose estimation as a unique optimisation problem tailored for CURL. Experiment evaluation shows that CURL-MAP achieves state-of-the-art 3D mapping results and competitive LiDAR odometry accuracy. We will release the CURL-MAP codes for the community.

The estimation of the 6D pose of industrial components is essential for smart manufacturing. Especially for complex units that require intensive manual operations, such as a concentrator photovoltaics solar panel, accurate spatial localization provides visual aids for industrial automation. In this paper, we propose an accurate and scalable framework to address the dimensional variability of industrial components and tackle practical implementation issues. First, we use the scalable architecture EfficientNet as the backbone coupled with an enhanced feature pyramid network to estimate the object’s pose. By introducing vertical and horizontal connections of shallow layers, the feature extraction of small objects is optimized for better detection accuracy. Second, leveraging the reliable 2D detection results and geometry information, an adaptive pose refinement module is designed to adjust the estimated 6D pose. The scaling of the backbone network and the computational complexity of refined modules are uniformly adjusted via a shared hyperparameter, resulting in a globally scalable framework. In terms of the pose estimation accuracy, the effectiveness of the refinement module and the real-time performance, validations are conducted both on the LINEMOD dataset and our customized datasets comprising of objects from the industrial photovoltaic system. Additionally, to further illustrate the effectiveness of the proposed method, a precision parallel robot is employed to validate the accuracy of real-time object pose tracking.

The assembly of 3D products with complex geometry and material, such as a concentrator photovoltaics solar panel unit, is typically conducted manually. This results in low efficiency, precision and throughput. This study is motivated by an actual industrial need and targeted towards automation of the currently manual assembly process. By replacing the manual assembly with robotic assembly systems, the efficiency and throughput could be improved. Prior to assembly, it is essential to estimate the pose of the objects to be assembled with high precision. The choice of the machine vision is important and plays a critical role in the overall accuracy of such a complex task. Therefore, this work focuses on the 6D pose estimation for precision assembly utilizing a 3D vision sensor. The sensor we use is a 3D structured light scanner which can generate high quality point cloud data in addition to 2D images. A 6D pose estimation method is developed for an actual industrial solar-cell object, which is one of the four objects of an assembly unit of concentrator photovoltaics solar panel. The proposed approach is a hybrid approach where a mask R-CNN network is trained on our custom dataset and the trained model is utilized such that the predicted 2D bounding boxes are used for point cloud segmentation. Then, the iterative closest point algorithm is used to estimate the object's pose by matching the CAD model to the segmented object in point cloud.

This paper proposes a novel design of a robotic hand exoskeleton device (PMHand) for the purpose of aiding post stroke rehabilitation. The main effects of a stroke on the human hand and the current rehabilitation methods and their limitations are briefly reviewed. The design process and fabrication of a full hand exoskeleton, control system and preliminary experimental results are presented in detail.

PKM (parallel kinematic machine) based production systems have shown their potential for large volume manufacturing. However, integrating 5-axis PKM machine tools directly into an existing system will lead to unnecessary redundant motions, which may cause many negative effects on the accuracy and productivity of the system. To cope with this problem, a new PKM based production system for aircraft wing assembly, named PAW, is proposed by using a holistic design approach in this paper. PAW is constructed by a unique PKM architecture mounted on a gantry-like worktable. Mobility, kinematics and Jacobian analyses are conducted, as well as dimensional optimization for the new PKM. Specially, the task space variables are directly employed for kinematic analysis, and the task workspace is directly utilized for the dimensional synthesis. As a result, the PAW system shows a better performance comparing to the those currently in production. It is not only suitable for wing assembly tasks (e.g. drilling) but also for other large volume manufacturing tasks, such as trimming and sealing.

This paper is a contribution to an invited session on the benchmarking of parallel mechanisms. The aim of the session is to compare different existing designs and prototypes of parallel mechanisms using a common set of benchmarking criteria. First, the kinematic architectures of parallel mechanisms of the multipteron family are presented. In addition to the tripteron and the quadrupteron, the pentapteron, a five-degree-of-freedom (dof) parallel mechanism is introduced. Then, the benchmarking criteria are applied to the prototypes of the tripteron (3-dof) and the quadrupteron (4-dof) prototypes. Although the tripteron and quadrupteron parallel mechanisms have been presented elsewhere, their properties, highlighted by the benchmarking analysis presented here are revealed for the first time.

3T1R four-degrees-of-freedom (DOF) parallel manipulators (3T1R-PMs) are the parallel counterparts of the 4-DOF SCARA serial robots. In a 3T1R-PM, the moving platform can generate 3T1R motion (also called Schonflies motion), which refers to a rotation about any axis with a given direction in conjunction with 3-DOF translations. A method is proposed for the type synthesis of 3T1R-PMs based on screw theory. The wrench systems of a 3T1R parallel kinematic chain (3T1R-PKC) and its legs are first analyzed. A general procedure is then proposed for the type synthesis of 3T1R-PMs. The type synthesis of legs for 3T1R-PKCs, the type synthesis of 3T1R-PKCs, as well as the selection of actuated joints of 3T1R-PMs, are dealt with in sequence. 3T1R-PKCs with and without inactive joints are synthesized. The phenomenon of dependent joint groups in a 3T1R-PKC is revealed for the first time. Several 3T1R-PMs with identical type of legs are obtained.