Robotics and Computer-integrated Manufacturing最新文献

Freeform surface parts play a significant role in the aerospace industry, the mold- manufacturing industry and the automobile industry, and it is energy-saving, material-saving, time-saving and environmentally beneficial to remanufacture the damaged components to restore their functionality and performance. Due to the complex geometry of the freeform surface wear, the adaptive remanufacturing of freeform surface parts is confronted with challenges. In this paper, an adaptive remanufacturing method for freeform surface parts based on linear laser scanner and robotic laser cladding is proposed to realize the precise freeform surface measurement and optimized remanufacturing path generation. On the one hand, a systematic wear measurement and assessment method is proposed to precisely locate and quantify the wear. With the noncontact calibration of the laser scanner and industrial robot, the contour of the target surface is real-timely measured and the reverse model is efficiently constructed, which provides detailed 3D morphological information of the worn freeform surface for the latter wear analysis. Next, considering the considerable difference between the reverse model and the nominal model, a refined model aligning method weighted by surface wear segmentation is proposed to minimize the alignment error and, further, the difference entity to be additively manufactured is obtained by discrete model comparison. On the other hand, to cope with the unsatisfactory binding strength over the freeform surface basis and small fragments of the working path for the traditional plane or cylinder slicing method, a novel remanufacturing path generation method is proposed. Considering the curvature distribution of the freeform surface, an optimized equidistant freeform surface slicing method is especially proposed for the difference entity to realize the adaptive fitting to the freeform basin. Furthermore, based on the equivalent volume overlapping model of laser cladding, the cladding track filling method for the freeform surface slicing is designed with the optimized track-to-track distance, which can reduce surface waviness and improve remanufacturing efficiency. Finally, simulations and experiments for the remanufacturing scenario of the steam turbine blade are conducted to verify the validity and feasibility of the proposed adaptive remanufacturing method for freeform surface parts based on linear laser scanner and robotic laser cladding.

Human-centric assembly is emerging as a promising paradigm for achieving mass personalization in the context of Industry 5.0, as it fully capitalizes on the advantages of human flexibility with robot assistance. However, in small-batch and highly customized assembly tasks, frequently changes in production procedures pose significant cognition challenges. To address this, leveraging computer vision technology to enhance human cognition becomes a feasible solution. Therefore, this review aims to explore the cognitive characteristics of human beings and classify existing computer vision technologies in a manner that discusses the future development of cognition-augmented human-centric assembly. The concept of cognition-augmented assembly is first proposed based on the brain's functional structure - the frontal, parietal, temporal, and occipital lobes. Corresponding to these brain regions, cognitive issues in spatiality, memory, knowledge, and decision-making are summarized. Recent studies conducted between 2014 and 2023 on visual computation of assembly are categorized into four groups: position registration, multi-layer recognition, contextual perception, and mixed-reality fusion, all aimed at addressing these cognitive challenges. The applications and limitations of current computer vision technology are discussed. Furthermore, considering the rapidly evolving technologies such as the metaverse, cloud services, large language models, and brain-computer interfaces, future trends on computer vision are prospected to augment human cognition corresponding to the cognitive issues.

For dynamic collision-free trajectory planning in dual-robot and human collaborative tasks, this paper develops an online dual-robot Mutual Collision Avoidance (MCA) scheme based on convex optimization. A novel convex optimization formulation model, named Disciplined Convex programming by Shifting reference paths (DCS), is proposed for solving the single-robot trajectory optimization problem. Furthermore, a new dual-robot trajectory convex optimization algorithm is presented for online adjustment of the dual-robot trajectories according to the collaborative task priority. The overall pipeline, named DCS-MCA, generates collision-free and time-optimal dual-robot trajectories, while prioritizing the task accessibility of the high-priority robot. Simulation experiments demonstrate that DCS exhibits comparable performance to the current state-of-the-art single-robot motion planner, while the DCS-MCA outperforms common algorithms by up to 30% in time optimality for dual-robot collaborative tasks. The feasibility and dynamic performance of the proposed approach are further validated in a real collaborative cell, illustrating its suitability for collaborative dual-robot tasks in moderately dynamic environments.

In light of the global economic downturn and the intricate division of labor in manufacturing, the imperative for advanced manufacturing services and Manufacturing Service Collaboration (MSC) has escalated significantly. As manufacturing services gravitate towards aggregation on manufacturing service platforms, platform-aggregated MSC has emerged as an inevitable and compelling trend, capturing the attention of researchers worldwide. However, despite the existence of numerous frameworks, models, operational mechanisms, and algorithms proposed for the platform-aggregated MSC, drawing comparisons between these studies remains a complex endeavor. To address this predicament, this article proposes and explores a novel benchmarking methodology for platform-aggregated MSC. By employing complex network theory, a comprehensive model of platform-aggregated MSC is constructed and supplemented with corresponding methodologies for data generation and the configuration of optimization algorithms. Moreover, pertinent performance evaluation metrics are scrutinized to assess their applicability in the context of platform-aggregated MSC. The article culminates with the execution of a series of platform operation experiments designed to test the effectiveness and practicality of the proposed benchmarking system, thereby contributing to the ongoing evolution of the MSC domain.

The concept of Operator 4.0 has been recently defined to evolve the modern industrial scenarios by defining a knowledge sharing process from/to operators and industrial systems, creating personalized skills, and introducing digital tools towards socially sustainable factories. In this context, dynamic and adaptive user interfaces can make humans part of the intelligent factory system, supporting human work contextually and providing specific contents when needed, preserving the human wellbeing. This paper defines a human-centric methodology for the symbiotic co-evolution of operators’ skills, assistive digital tools and user interfaces, developed within the Horizon Europe project titled “DaCapo - Digital assets and tools for Circular value chains and manufacturing products”. The project focuses on defining a new set of human-centric digital tools and services for the manufacturing industry capable of boosting the application of circular economy (CE) throughout the manufacturing value chains. The proposed methodology can link the specific needs of an industrial case to the definition of the most proper assistive digital tools and functionalities to drive the design of adaptive, proactive user interfaces for the Operator 4.0. The method has been applied and validated on one of the project use cases, involving a manufacturing company operating in warehousing and logistics.

This paper addresses an integrated rack assignment and robot routing problem arising in robotic movable fulfillment systems (RMFS). This NP-hard planning task goes beyond current literature by simultaneously optimizing movable rack selection and multi-agent collision-free path finding, rather than decomposing them. A mixed integer programming (MIP) model with a new level-space-time network representation is proposed, jointly considering reusable racks, robot-rack pairings, storage repositioning, and collision avoidance. To improve computational efficiency, a fast rolling horizon heuristic and greedy algorithm are developed. Extensive experiments demonstrate that the integrated method's solutions can improve by 30 % upon conventional decomposed approaches. Intriguing test cases reveal the model, suggesting non-intuitive robot carryover policies that are unfound by separate selection and routing methods. This indicates potential optimization benefits from explicitly coordinating task assignment, scheduling, and routing decisions in complex automated warehousing systems. The rolling horizon heuristic solutions approach optimality with much greater efficiency than directly solving one large MIP, validating its practical value. This research provides useful integrated modeling insights, efficient solution algorithms, and decision support for efficiently controlling next-generation robotic movable fulfillment systems.

On the landscape of solutions to deal with delicate objects, the development and use of soft grippers is a topic of increasing interest, with a large number of prototypes proposed by the research community employing non-linear soft materials and based on diverse actuation means. However, increasing compliance usually leads to the reduction of lifting capacity. As a recent promising approach, shear forces exerted by a soft gripper can be enhanced by exploiting the electro-adhesion (EA) effect. Following this research trend, this paper proposes a new gripper that combines a compliant finger structure, with geometry taken from the FESTO FinRay but made of a softer material (a urethane rubber), and custom EA pads that are placed on the fingers at the interface with the grasped object. Following hyper-elastic model identification of the considered material and preliminary functional verification of gripper design via finite element simulations, the gripper is then manufactured and tested by means of a specific setup, replicating the grasping and lifting of cylindrical objects with different diameters. The results clearly show that the new gripper makes it possible to generate holding forces similar to those of the FESTO FinRay, but with significantly lower pressures on the grasped object (77 % less). Besides enabling the handling of more fragile items, the drastic increase in gripper compliance also results in lower mechanical actuation force (namely, 71 % less of gripping energy) required to generate the same holding force, with a consequent reduction of operation costs and sustainability of its application.

In sheet metal forming, the quality of a formed part is strongly influenced by the local lubrication conditions on the blank. Fluctuations in lubrication distribution can cause failures such as excessive thinning and cracks. Predicting these failures in real-time for the entire part is still a very challenging task. Machine learning (ML) based digital twins and advanced computing power offer new ways to analyze manufacturing processes inline in the shortest possible time. This study presents a digital twin for simulating a deep drawing process that incorporates an advanced ML model and optimization algorithm. Convolutional neural networks with RES-SE-U-Net architecture, were used to capture the full friction conditions on the blank. The ML model was trained with data from a calibrated finite element model. The ML model establishes a correlation between the local friction conditions across the blank and the quality of the drawn part. It accurately predicts the geometry and thinning of the formed part in real-time by assessing the friction conditions on the blank. A particle swarm optimization algorithm incorporates the ML model and provides tailored recommendations for adjusting local friction conditions to promptly correct detected quality deviations with minimal amount of additional lubricant. Experiments show that the ML model deployed on an industrial control system can predict part quality in real-time and recommend adjustments in case of quality deviation in 1.6 s. The error between prediction and ground truth is on average 0.16 mm for geometric accuracy and 0.02 % for thinning.

Facing the new generation intelligent manufacturing, traditional manufacturing models are transitioning towards large-scale customized productions, improving the efficiency and flexibility of complex manufacturing processes. This is crucial for enhancing the stability and core competitiveness of the manufacturing industry, and human-robot collaboration systems are an important means to achieve this goal. At present, mainstream manufacturing human-robot collaboration systems are modeled for specific scenarios and actions, with poor scalability and flexibility, making it difficult to flexibly handle actions beyond the set. Therefore, this article proposes a new human-robot collaboration framework based on action recognition and multi-scale control, designs 27 basic gesture actions for motion control, and constructs a robot control instruction library containing 70 different semantics based on these actions. By integrating static gesture recognition, dynamic action process recognition, and You-Only-Look-Once V5 object recognition and positioning technology, accurate recognition of various control actions has been achieved. The recognition accuracy of 27 types of static control actions has reached 100%, and the dynamic action recognition accuracy of the gearbox assembly process based on lightweight MF-AE-NNOBJ has reached 90%. This provides new ideas for simplifying the complexity of human-robot collaboration problems, improving system accuracy, efficiency, and stability.

In the process of metal cutting, realizing effective monitoring of tool wear is of great significance to ensure the quality of parts machining. To address the tool wear monitoring (TWM) problem, a tool wear monitoring method based on data-driven and physical output is proposed. The method divides two Physical models (PM) into multiple stages according to the tool wear in real machining scenarios, making the coefficients of PM variable. Meanwhile, by analyzing the monitoring capabilities of different PMs at each stage and fusing them, the PM's ability to deal with complex nonlinear relationships, which is difficult to handle, is improved, and the flexibility of the model is improved; The pre-processed signal data features were extracted, and the original features were fused and downscaled using Stacked Sparse Auto-Encoder (SSAE) networker to build a data-driven model (DDM). At the same time, the DDM is used as a guidance layer to guide the fused PM for the prediction of wear amount at each stage of the tool, which enhances the interpretability of the monitoring model. The experimental results show that the proposed method can realize the accurate monitoring of tool wear, which has a certain reference value for the flexible tool change in the actual metal-cutting process.