Car body design includes the construction of car sketches, 2D models and 3D models. The establishment of high-quality car body models usually relies on CAD software, which requires the proficient operating skills, and the efficiency is relatively low compared with automatic modeling. The time cost of communication between the designer and the model engineer will also affect the modeling efficiency. How to quickly and automatically generate high quality parametric models from car sketches or images efficiently has become a development direction for the automotive design. In this paper, we propose a template-based reconstruction method of 2D high-precision car body model. On the basis of the established 2D model database, a coarse model is achieved by the improved Orthogonal Matching Pursuit algorithm with few key points which are obtained by deep neural network; further, according to the proposed auto-fitting optimization algorithm of cubic Bezier curve, the modeling process of "from coarse to fine" is realized combining edge information. The proposed template-based 2D high-precision model reconstruction algorithm of the car body can greatly reduce the modeling time under the given modeling accuracy.

The goal of the PITSTOP (Immersive platform for structured operator training) project is to overcome the limits of traditional workplace training on dangerous systems and reduce the related risks, using an innovative integration of engineering simulation models and virtual reality (VR) tools. This article specifically presents the first VR platform for training on small-scale industrial steam generators, representative of a vast class of hazardous industrial equipment. The dynamic model of the steam generator was developed in matlab-simulink using a mixed physics-based and data-driven approach. The generator model includes the main engineering components, actuators and measuring equipment, as well as control logic and emergency procedures. It can simulate normal operations and emergency situations. The model was calibrated using experimental data collected from the real system at various operating conditions to align simulated performance with the real behavior. The VR environment was developed in Unity, a graphics engine widely adopted by the videogame industry, using three-dimensional computer-aided design models of the steam generator and its surroundings. The user can access this immersive system wearing an HTC Vive headset. Unlike most existing training systems, learners can interact with the actuators using bare hands gestures, without controllers, making the experience intuitive and easily accessible. By connecting the dynamic model with the VR environment, the user's interactions are directly provided to the steam generator model, which in turn directly outputs the steam generator response to the VR environment, providing audio and visual feedback to the user, as if they were actually acting on the real generator. The results from this study could boost the further development of training platforms to safely train operators and certify their competence.

The work here presented is part of a wider research project aimed at extracting and using in industrial applications high level semantic information from 3D product models that are described by means of their boundary representation (B-rep). The specific focus of the paper is the recognition among the components of the CAD model of an assembly of those belonging to some categories of standard parts largely employed in mechanical industry. The knowledge of these components is crucial to understand the structure of mechanical products as they have specific meaning and function. Standard parts follow international standard in shape and dimensions, and also typical mounting schemes exist concerning their use in the product assembly. These distinctive features have been exploited as a starting point to develop a multi-step recognition algorithm. It includes a shape-based and a context-based analysis both relying on the geometric and topological analysis of a CAD model. As already anticipated by Voelcker in his visionary ability to anticipate open challenges, the shape of an object alone is not enough to understand its function. Therefore, context assessment becomes crucial to validate the recognition given by the shape-based step. It allows to uniquely recognize components in mechanical CAD models, by confirming correct results, refusing the false positives, as well as choosing the correct one when the assignment is multiple.

The use of VR immersive systems for the training of operators in various industrial contexts is becoming more and more common thanks to the possibility of simulating a great variety of situations in safe conditions, which can hardly be experienced in reality without risks. One of the main issues in the development of such training systems is the cost for the creation of alternative learning scenarios, which normally requires the simultaneous involvement of the training expert for its design and the IT specialist for its realization in the VR environment. To reduce this effort, this paper proposes a methodology to allow the training expert to directly specify the training scenarios using familiar tools. The method is devised considering the training of steam generator operators, but it can be easily adapted for other industrial equipment.

Design For Assembly (DFA) aims at improving product design facilitating assembly phases via the application of evaluation metrics and design guidelines. However, DFA analyses are usually performed manually and the adoption of supporting tool is poor. This paper investigates the application of algorithms allowing to extract from CAD assembly models the required data to perform automated DFA analyses, thus providing a tool to support designers' everyday works. In particular, attributes from geometric feature recognition algorithms, solids properties and assembly parts' semantics are leveraged and mapped to the parameters required to accomplish DFA evaluations. The proposed approach is illustrated on a 3D printer for home use. At first, a manual DFA analysis has been performed on the product identifying product BOM, components properties, assembly cycle and times according to models in the literature. Then, the CAD model of the printer has been processed with some geometric algorithms to verify the possibility to extract the required data to be used as input to the DFA analysis. The test case has demonstrated the feasibility of the approach, even if some design considerations and improvement directions still need the critical evaluation of the designer.

N/A

N/A

Over the past decades, attention towards workplace safety has increased progressively, leading to a strong presence of automation and remote control on hazardous processes and devices. However, human intervention is sometimes necessary. Consequently, inadequate knowledge of equipment, poor maintenance of work tools and underestimation of possible risks can result into work accidents. Therefore, it is fundamental for each worker to be updated about necessary knowledge to run and maintain potentially dangerous work equipment. This knowledge is generally acquired through both theoretical studies and training on real devices, sometimes leading to the acquisition of a specific licence. However, this kind of training can test the ability to run the real system only during its standard behavior, but not during emergency situations or malfunctions. Moreover, the interaction of an unexperienced operator with a potentially hazardous system during the training process can lead to risky scenarios. The goal of the PITSTOP project is to overcome the limits of traditional training and to reduce the possible risks of practical tests, thanks to an innovative integration of dynamic simulation models and a virtual reality environment. In this article, the case study of a small-scale steam generator for industrial applications is considered. However, the methodology proposed by the PITSTOP project could be easily extended to other hazardous systems. A dynamic model of the system was developed in Matlab-Simulink, including all the main devices (i.e., water tank, pump, boiler), actuators (i.e., valves, buttons, switches, levers) and measuring equipment (i.e. temperature probes, pressure sensors, etc). The simulation of the steam generator relies on a mixed physics-based/data-driven approach, based on physical equations, semi-empirical correlations, performance maps and a thermophysical properties tool (Coolprop). The model was designed to simulate the normal operation of the system during stationary and transient scenarios, but also to recreate emergency situations and to show the effect of wrong or inconsistent actions by the operator. The control logics and emergency procedures of the steam generator were included in the model as well. Operating the real system in various conditions, it was possible to collect experimental data characterizing its behavior, and to understand all the possible interactions of the operator with its actuators and their consequences. Experimental data were used to calibrate the Matlab-Simulink component models and to align the performance of the model with the real steam generator. The virtual reality environment was developed in Unity, a graphics engine widely adopted by the videogame industry, using 3D CAD models of the steam generator and its surroundings. The user can access to this immersive system wearing an HTC VIVE headset. No other equipment must be worn, making the experience intuitive and easily accessible. In fact, the movement of the user in the environment is detected by two HTC VIVE cameras installed inside the testing room, while the interaction in the virtual reality environment is guaranteed by hand gestures that are captured and interpreted using Leap Motion controller. Connecting the dynamic model with the virtual reality environment via UDP communication, it is possible to reproduce faithfully the interactions with the steam generator. The actions of the user on the actuators are used as inputs of the model. Simulating the system response over a time step, the outputs of the model (e.g., temperatures and pressures) are sent to the virtual reality environment, providing visual and audio feedbacks to the user.

With the widespread use of sensors able to track human motions and gestures, natural interactions are becoming pervasive in many application contexts. The design of such interactions depends on the precision of the tracking device as well as on the background of the user and his/her level of confidence in such technologies. In this paper, we present a set of 3D interaction techniques where the user does not need to wear any additional device for hand tracking, and they can interact in the virtual environment with his/her bare hands directly. The usability and the naturalness of the proposed interaction have been tested on a case study where users interact with a steam generator for training purposes. (Abstract)

Immersive virtual reality systems allow simulating realistic situations in which the user can learn the use of dangerous machinery through structured paths and actively learn how to manage dangerous situations in total safety. The use of VR for training of industrial equipment while not new is not as widespread due to the cost of creating and adapting it to address various equipment and situations. To overcome this limitation, this paper proposes a VR simulator for training steam generator operators and verifiers focusing on the easy of customizing the VR system to suit new learning paths and different equipment while reducing the implementation efforts by ITC experts.

Due to its capacity to evolve in a large solution space, the Simulated Annealing (SA) algorithm has shown very promising results for the Reverse Engineering of editable CAD geometries including parametric 2D sketches, 3D CAD parts and assemblies. However, parameter setting is a key factor for its performance, but it is also awkward work. This paper addresses the way a SA-based Reverse Engineering technique can be enhanced by identifying its optimal default setting parameters for the fitting of CAD geometries to point clouds of digitized parts. The method integrates a sensitivity analysis to characterize the impact of the variations in the parameters of a CAD model on the evolution of the deviation between the CAD model itself and the point cloud to be fitted. The principles underpinning the adopted fitting algorithm are briefly recalled. A framework that uses design of experiments (DOEs) is introduced to identify and save in a database the best setting parameter values for given CAD models. This database is then exploited when considering the fitting of a new CAD model. Using similarity assessment, it is then possible to reuse the best setting parameter values of the most similar CAD model found in the database. The applied sensitivity analysis is described together with the comparison of the resulting sensitivity evolution curves with the changes in the CAD model parameters imposed by the SA algorithm. Possible improvements suggested by the analysis are implemented to enhance the efficiency of SA-based fitting. The overall approach is illustrated on the fitting of single mechanical parts but it can be directly extended to the fitting of parts' assemblies. It is particularly interesting in the context of the Industry 4.0 to update and maintain the coherence of the digital twins with respect to the evolution of the associated physical products and systems.

This paper introduces a novel reverse engineering technique for the reconstruction of editable CAD models of mechanical parts' assemblies. The input is a point cloud of a mechanical parts' assembly that has been acquired as a whole, i.e. without disassembling it prior to its digitization. The proposed framework allows for the reconstruction of the parametric CAD assembly model through a multi-step reconstruction and fitting approach. It is modular and it supports various exploitation scenarios depending on the available data and starting point. It also handles incomplete datasets. The reconstruction process starts from roughly sketched and parameterized geometries (i.e 2D sketches, 3D parts or assemblies) that are then used as input of a simulated annealingbased fitting algorithm, which minimizes the deviation between the point cloud and the reconstructed geometries. The coherence of the CAD models is maintained by a CAD modeler that performs the updates and satisfies the geometric constraints as the fitting process goes on. The optimization process leverages a two-level filtering technique able to cap- ture and manage the boundaries of the geometries inside the overall point cloud in order to allow for local fitting and interfaces detection. It is a user-driven approach where the user decides what are the most suitable steps and sequence to operate. It has been tested and validated on both real scanned point clouds and as-scanned virtually generated point clouds incorporating several artifacts that would appear with real acquisition devices

Focus of this work is the recognition of the standard parts contained in a CAD assembly model, with the aim of enhancing the model semantics. Standard parts are components typically used in mechanical industry, which have a specificc engineering meaning and follow international standards. In particular eight categories of standard parts are considered, i.e. screws, nuts, O-ring, washers, circlips,keys, studs and pins. The provided algorithm relies on the geometric and topological analysis of the CAD model parts. A part is assigned to one of the categories if it satisfies the geometric requirements extracted for that specific category, based on engineering knowledge and design rules. In addition, if a part is recognized as standard part, besides the class of membership, further information is provided as result, namely its engineering dimensions.

CAD3A (CAD Assembly Analisys Application), a web-based application specifically designed to cope with assembly models. CAD3A allows the visualization of standard CAD models as well as the computation and the visualization of their salient characteristics implicitly embedded. In particular, it allows (i) detecting the degree of freedom (DOF) of each pair of parts sharing a surface contact and (ii) computing a functional classification of the parts.

In the last decades, digital 3D models have substituted 2D technical drawings for the design and development of mechanical products. This results in a huge amount of CAD (computer-aided design) models available in legacy and Web repositories. 3D CAD models are now the reference product specification, and are frequently exchanged with customers and companies collaborating in the product development. In addition, this great amount of data can be exploited by scientists to test and evaluate their research results. Therefore, the visualization and analysis of CAD models from everywhere and through any device are particularly useful for both engineers and researchers. To answer this need, this paper presents a web-based solution exploiting the X3D capabilities for the direct visualization in the Web browser of CAD assemblies and of some of their characteristics remotely computed by the application on user demand.

n/a

Gesture recognition is a fundamental tool to enable novel interaction paradigms in a variety of application scenarios like Mixed Reality environments, touchless public kiosks, entertainment systems, and more. Recognition of hand gestures can be nowadays performed directly from the stream of hand skeletons estimated by software provided by low-cost trackers (Ultraleap) and MR headsets (Hololens, Oculus Quest) or by video processing software modules (e.g. Google Mediapipe). Despite the recent advancements in gesture and action recognition from skeletons, it is unclear how well the current state-of-the-art techniques can perform in a real-world scenario for the recognition of a wide set of heterogeneous gestures, as many benchmarks do not test online recognition and use limited dictionaries. This motivated the proposal of the SHREC 2021: Track on Skeleton-based Hand Gesture Recognition in the Wild. For this contest, we created a novel dataset with heterogeneous gestures featuring different types and duration. These gestures have to be found inside sequences in an online recognition scenario. This paper presents the result of the contest, showing the performances of the techniques proposed by four research groups on the challenging task compared with a simple baseline method.

Mechanical assemblies are very complex structures, made of many parts of various shapes and sizes with different usages. Consequently, it ischallenging to manage them during all the manufacturing processes, from the design to the assembly and the recycling. Aiming to simplify theassembly structure and reduce the number of parts to deal with simultaneously, in literature many works exist on subassemblies identificationstarting from the CAD assembly model. However, the methods provided loose sight of many details associated with the parts, as well as the factthat the treated model represents a real mechanical assembly which respects precise engineering rules. At this regard, this work introduces a novelmethodology to detect meaningful clusters in CAD assembly models. The logic applied relies on engineering knowledge, both of mechanicalassemblies' components and of assembling techniques, and on the leveraging of the semantics of components. In particular, referring to generaldesign rules, we have identified some heuristics to exploit to partition the assembly into different types of clusters, such as the symmetry alongan axis and the presence of fasteners or welds. It results that the assembly's parts are meaningfully grouped, considering, at the same time, theirshape, functionality, and type of contact.

In industrial manufacturing, both in the design and the production phase, the management of modern mechanical assemblies is becoming demanding due to their increasing complexity. The use of stable subassemblies concept constitutes a better alternative, which allows to independently treat smaller groups of the assembly's parts, also to achieve a parallel production. At this regard, several methods for automatic subassemblies identifi-cation, starting from the assembly CAD model, have been provided. However, most of the methodologies proposed rely on human intervention, especially in the model processing to make available essential data, while other details are ignored. After giving the definition of stable subassembly, this paper focuses on the application of stable subassemblies identification to industrial CAD models and highlights the issues arising. With the aim of ensuring a reliable CAD model analysis, starting point of the identification, the possible real engineering situations, both related to assembling methods and modelling techniques, are presented. Ap-proaches to algorithmically address them are then described, with the help of two examples of mechanical assemblies.

This paper addresses the way a simulated annealing-based fitting strategy can be enhanced by leveraging a sensitivity analysis able to characterize the impact of the variations in the parameters of a CAD model on the evolution of the deviation between the CAD model itself and the point cloud of the digitized part to be fitted. The principles underpinning the adopted fitting algorithm are briefly recalled. The applied sensitivity analysis is described together with the comparison of the result-ing sensitivity evolution curves with the changes in the CAD model parameters im-posed by the simulated annealing algorithm. This analysis suggests several possible improvements that are discussed. The overall approach is illustrated on the fitting of single mechanical parts but it can be directly extended to the fitting of parts' assemblies. It is particularly interesting in the context of the Industry 4.0 to update digital twins of physical products and systems.