Search results

For robots to more actively interact with the surrounding environment in object manipulation tasks or walking, they must understand the physical attributes of objects and surface materials they encounter. Dynamic tactile sensing can effectively capture rich information about material properties. Hence, methods that convey and interpret this tactile information to the user can improve the quality of human–machine interaction. This paper aims to propose a visual-tactile cross-modal retrieval framework to convey tactile information of surface material for perceptual estimation.

The tactile information of a new unknown surface material can be used to retrieve perceptually similar surface from an available surface visual sample set by associating tactile information to visual information of material surfaces. For the proposed framework, the authors propose an online low-rank similarity learning method, which can effectively and efficiently capture the cross-modal relative similarity between visual and tactile modalities.

Experimental results conducted on the Technischen Universität München Haptic Texture Database demonstrate the effectiveness of the proposed framework and the method.

This paper provides a visual-tactile cross-modal perception method for recognizing material surface. By the method, a robot can communicate and interpret the conveyed information about the surface material properties to the user; it will further improve the quality of robot interaction.

The purpose of this study is to develop an output model to extract surface microstructure characteristics of different objects, so as to predict the response of the output voltage obtained from tactile texture sensor.

The model is based on the consideration of the inverse-magnetostrictive effect, the flexure mode, the linear constitutive equations and the strain principle.

This research predicts and investigates the effect of the texture properties on the tactile texture sensor output characteristics.

The surface texture characteristic is regarded to be important information to evaluate and recognize the object.

Automatic defect detection is a fundamental and vital topic in the research field of industrial intelligence. In this work, the authors develop a more flexible deep learning method for the industrial defect detection.

The authors propose a unified framework for detecting defects in industrial products or planar surfaces based on an end-to-end learning strategy. A lightweight deep learning architecture for blade defect detection is specifically demonstrated. In addition, a blade defect data set is collected with the dual-arm image collection system.

Numerous experiments are conducted on the collected data set, and experimental results demonstrate that the proposed system can achieve satisfactory performance over other methods. Furthermore, the data equalization operation helps for a better defect detection result.

An end-to-end learning framework is established for defect detection. Although the adopted fully convolutional network has been extensively used for semantic segmentation in images, to the best knowledge of the authors, it has not been used for industrial defect detection. To remedy the difficulties of blade defect detection which has been analyzed above, the authors develop a new network architecture which integrates the residue learning to perform the efficient defect detection. A dual-arm data collection platform is constructed and extensive experimental validation are conducted.

The purpose of this paper is to propose a three-dimensional (3D) assembly model retrieval method based on assembling semantic information to address semantic mismatches, poor accuracy and low efficiency in existing 3D assembly model retrieval methods.

The paper proposes an assembly model retrieval method. First, assembly information retrieval is performed, and 3D models that conform to the design intention of the assembly are found by retrieving the code. On this basis, because there are conjugate subgraphs between attributed adjacency graphs (AAG) that have an assembly relationship, the assembly model geometric retrieval is translated into a problem of finding AAGs with a conjugate subgraph. Finally, the frequent subgraph mining method is used to retrieve AAGs with conjugate subgraphs.

The method improved the efficiency and accuracy of assembly model retrieval.

The examples illustrate the specific retrieval process and verify the feasibility and reasonability of the assembly model retrieval method in practical applications.

The assembly model retrieval method in the paper is an original method. Compared with other methods, good results were obtained.

The purpose of this study is the use of 3D printing technology to perform maintenance on damaged parts on site. To maintain damaged parts, the user needs experience in the parts design and 3D printing technology. To help users who have little or no experience on 3D printing, a part library-based information retrieval and inspection framework was proposed to support the process of manufacturing replaceable parts using a 3D printer.

To establish the framework, 3D printing-based maintenance procedure was first defined, comprising retrieval, manufacturing and inspection steps, while identifying the technical components required to perform the procedure. Once the technical components are identified, part library-based information retrieval and inspection framework was defined based on the technical components and the relationships between the components. For validation of the concept of the framework, prototype system is developed according to the proposed framework.

The feasibility of the proposed framework is proved through maintenance experiments on gaskets and O-rings.

The main contribution of this study is the proposal of the framework, which aims to support the maintenance of damaged parts for the user who has little or no experience in part design or does not know how to operate a 3D printer.

The purpose of this paper is to introduce a novel technique for printing with multiple materials using the DLP method. Digital-light-processing (DLP) printing uses a digital projector to selectively cure a full layer of resin using a mask image. One of the challenges with DLP printing is the difficulty of incorporating multiple materials within the same part. As the part is cured within a liquid basin, resin switching introduces issues of cross-contamination and significantly increased print time.

The material handling challenges are investigated and addressed by taking inspiration from automated storage and retrieval systems and using an active cleaning solution. The material tower is a compact design to facilitate the storage and retrieval of different materials during the printing process. A spray mechanism is used for actively cleaning excess resin from the part between material changes.

Challenges encountered within the multi-material DLP technology are addressed and the experimental prototype validates the proposed solution. The system has a cleaning effectiveness of over 90 per cent in 15 s with the build area of 72 inches, in contrast to the previous work of 50 per cent cleaning effectiveness in 2 min with only 6 inches build area. The method can also hold more materials than the previous work.

The techniques from automated storage and retrieval system is applied to develop a storage system so that the time complexity of swapping is reduced from linear to constant. The whole system is sustainable and scalable by using a spraying mechanism. The design of the printer is modular and highly customizable, and the material waste for build materials and cleaning solution is minimized.

A fixture is a special tool used to accurately and stably locate the workpiece during machining process. Proper fixture design improves the quality and production of parts and also facilitates the interchangeability of parts, which is prevalent in much of modern manufacturing. The purpose of this paper is to combine the rule‐based reasoning (RBR) and case‐based reasoning (CBR) method for machining fixture design in a virtual reality (VR) based integrated system.

In this paper, an approach combining the RBR and fuzzy comprehensive judgment method is proposed for reasoning suitable locating features and clamping features. Based on the reasoning results, a CBR method for machining fixture design is then presented.

The paper finds that the proposed system is an efficient tool for machining fixture design.

The proposed system enables the designers to perform fixture design with automated fixture locating method reasoning and make a new fixturing solution quickly by referencing previous design cases.

A VR application for machining fixture design is presented.

Although the Boolean combination of keywords and/or subject codes is the predominant access method for the retrieval of passages from full‐text databases, menu access is an attractive alternative. The selection of an access method and the ensuing satisfaction with the results is based on the type of query and on the experience and knowledge of the user. This paper describes a prototype system which has integrated Boolean, menu, and direct access methods for the retrieval of passages from full‐text databases. The integration is based on the hierarchical structure inherent in such databases as legal statutes and regulations and engineering standards. The user may switch freely among access methods in order to develop the most appropriate search strategy. The retrieved passages are presented to the user within the context of the hierarchical structure.

This study aims to compute 3D model similarity by extracting and comparing shape features from the neutral files.

In this work, the clear text encoding document STEP (Standard for The Exchange of Product model data) of 3D models was analysed, and the models were characterized by two-depth trees consisting of both surface and shell nodes. All surfaces in the STEP files can be subdivided into three kinds, namely, free, analytical and loop surfaces. Surface similarity is defined by the variation coefficients of distances between data points on two surfaces, and subsequently, the shell similarity and 3D model similarity are determined using an optimal algorithm for bipartite graph matching.

This approach is used to experimentally verify the effectiveness of the 3D model similarity algorithm.

The novelty of this study research lies in the computation of 3D model similarity by comparison of all surfaces. In addition, the study makes several key observations: surfaces reflect the most information concerning the functions and attributes of a 3D model and so the similarity between surfaces generates more comprehensive content (both external and internal); semantic-based 3D retrieval can be obtained under the premise of comparison of surface semantics; and more accurate similarity of 3D models can be obtained using the optimal algorithm of bipartite graph matching for all surfaces.