1,721,133 research outputs found
Deep Learning-Based Method for Vision-Guided Robotic Grasping of Unknown Objects
Full text linkCollaborative robots must operate safely and efficiently in ever-changing unstructured environments, grasping and manipulating many different objects. Artificial vision has proved to be collaborative robots' ideal sensing technology and it is widely used for identifying the objects to manipulate and for detecting their optimal grasping. One of the main drawbacks of state of the art robotic vision systems is the long training needed for teaching the identification and optimal grasps of each object, which leads to a strong reduction of the robot productivity and overall operating flexibility. To overcome such limit, we propose an engineering method, based on deep learning techniques, for the detection of the robotic grasps of unknown objects in an unstructured environment, which should enable collaborative robots to autonomously generate grasping strategies without the need of training and programming. A novel loss function for the training of the grasp prediction network has been developed and proved to work well also with low resolution 2-D images, then allowing the use of a single, smaller and low cost camera, that can be better integrated in robotic end-effectors. Despite the availability of less information (resolution and depth) a 75% of accuracy has been achieved on the Cornell data set and it is shown that our implementation of the loss function does not suffer of the common problems reported in literature. The system has been implemented using the ROS framework and tested on a Baxter collaborative robot
Early sustainability assessment to design competitive industrial systems
No full textNowadays industrial system design has to face a big issue: offering new advanced functionalities, guaranteeing high performances, respecting the competitive pressure, limiting the environmental impact, expanding the company’s market share, being usable and easy to control. In a nutshell, they have to be sustainable in respect with planet, profit and people. In this context, Design for Sustainability (D4S) promotes a sustainable design practice, where all impacts are considered and optimized. However, D4S methods usually focus on one single aspect at a time (e.g. eco-design addresses environmental issues, ergonomics investigates physical human-product interaction, etc.). In practice, industrial systems design requires numerous aspects to be integrated and optimized contemporarily and interactively: mechanics, electronics, system control, management of material and information flows, human-machine interface, human-product interaction, as well as impacts on environment, costs and human factors. The present research proposes an analytical approach for an early sustainability assessment based on a set of Key Performance Indicators (KPIs) considering the three aspects of sustainability (environment, cost and humans) and a feature-based approach, to support their easy and preventive analysis
Computer aided engineering optimization of two screw based innovating intra-medullary fixation devices
No full textA new generation of intra-medullar fixation devices was designed and developed through intensive Computer Aided Engineering research studies. The target of the research activities was the development of a novel design of intra-medullar nails for pertrochanteric fractures with enhanced performances on the in vivo subjects
Digital Manufacturing and Virtual Reality for Tractors' Human-Centred Design
Full text linkHuman-centred design is based on the satisfaction of the user needs related to performances, aesthetics, reliability, usability, accessibility and visibility issues, costs, and many other aspects. The combination of all these aspects has been called as “perceived quality”, that is definitely a transdisciplinary topic. However, the “real” perceived quality is usually faithfully assessed only at the end of the design process, while it is very difficult to predict on 3D CAD model. In this context, digital manufacturing tools and virtual simulation technologies can be validly used according to a transdisciplinary approach to create interactive digital mock-ups where the human-system interaction can be simulated and the perceived quality assessed in advance. The paper proposes a mixed reality (MR) set-up where systems and humans interacting with them are digitalized and monitored to easily evaluate the human-machine interaction. It is useful to predict the design criticalities and to improve the global system design. An industrial case study has been developed in collaboration with CNH Industrial to demonstrate how the proposed set-up can be validly used to support human-centred design
Adaptive Manufacturing: Challenges to the Industrial and Scientific Community
No full textAdaptive manufacturing posesmany challenges to the industrial and scientific community. One of the main interesting issues, still requiring further research efforts, consist in achieving effective modularisation of both control systems and related physical machinery. Modularisation brings evident advantages towards effective and fast reconfiguration of assembly lines, maintaining at the same time high reliability of the single machinery as well as safety and effectiveness of more complex production units. Safety and reliability of either a single device or production cell are however not enough in order to ensure safety of more complex assembly units. Novel methodologies which consider manufacturing at the system level are then required. In general, safety (as well as hazard) is indeed an emergent systemic property and as such requires to be dealt with specific, system oriented, methodologies. Such methodologies allow to obtain control software which is “correct-by-design” and are tightly integrated with design: interestingly, such methodologies may be applied from the intra-device level to single devices, machinery, production units, assembly lines. As complexity increases, it becomes possible to deal with single primary faults within simple mechanical components (for example small engines and actuators) to more complex units, for example production cells and assembly lines, providing increasing levels of failure detection and protection, ranging from primary faults to fail operational and fail safe behaviours
Parametric virtual concepts in the early design of mechanical systems: a case study application
No full textVirtual prototyping enables the validation and optimization of mechanical devices similar to physical testing, saving time and costs in the product development, especially in case of heavy machines with complex motions. However, virtual prototyping is usually deployed only at the end of the design process, when the product architecture has already been developed. The present paper discusses the introduction of virtual prototypes since the conceptual design stage as “Virtual Concepts”, in which coarse models of machinery design variants are simulated obtaining useful information, sometimes fundamental to support best design choices. Virtual Concept modeling and preliminary validation, along with its later integration into a Virtual Prototype, are expressly investigated using Multi Body Dynamics software. A verification case study concerning a large vibrating screen is presented, in order to demonstrate that dynamic Virtual Concepts can enable an easier and effective evaluation of the design variants, thus increasing the design process predictability. Finally, current challenges to be solved for the practical adoption of Virtual Concept simulations as an integral part of the industrial design process are critically discussed
Design of Iron Cast Deburring Robotic Cells with Simulation and Offline Programming Tools
No full textThe design of last generation anthropomorphous robots manufacturing cells is a challenging activity, to be developed accurately in a fast and effective way due to the strong requirements to satisfy in terms of time to market, overall performances and the total costs of the final application.
During the design stages, different types of graphical and technical representations are needed to exalt and communicate the design intent and to evaluate the effects of every design variant on the final operating performances.
For this reasons the choice of the tools to adopt, and a correct methodology for an integrated use, for this purpose is strategically important and delicate. On one side the design tools must be focused to point out all the multidisciplinary effects involved, especially those concerning the strong interrelation of the mechanical apparatus solutions with the electrical and electronics systems, while on the other side they must be absolutely integrated and easy to use.
In the present paper it will be explained the choices effected on the graphical representation and design tools, and the design method created for an effective integrated use, in the design and development of iron casts deburring anthropomorphous robots manufacturing cells, performed during real experiences, developed together by DIMEC University researchers and SIR, an important system integrator enterprise.
Actually, state of the art 3D parametric CAD and behavioural simulation design tools confirmed to be powerful but the adoption of anthropomorphous robot simulation and offline programming tools proved to be fundamental for the final validation of every design variant.
The need to enrich classical mechanical design representation, static and fixed for single configurations, with animations where robot movements and the motion control can be effectively communicated and evaluated, was then solved and proved to give an important feedback to the design engineers, who acquired more consciousness on the real performances achievable.
Furthermore, the availability of multimedia movies, was finally appreciated even to interact with the final user for training purposes
How to analyse the workers’ experience in integrated product-process design
No full texthe analysis of workers’ ergonomics and human factors is assuming a great importance in product and process design for modern industry. However, there is a lack of common references and structured protocols for the assessment of workers’ experience in industrial practices in an effective and predictive way. As a result, designers are poorly supported in the application of digital technologies, which are demonstrating to have a great potential. This ascertainment suggested defining a reference model to analyse the so-called user experience (UX) of workers and a proper technological set-up based on virtual simulations in order to support human-centred product-process design. Indeed, the recent advances in ubiquitous computing, wearable technologies and low-cost connected devices offer a huge amount of new tools for human data monitoring. However, the open issue is selecting the most proper devices for industrial application area in respect with design goals, using virtual simulation and digital manufacturing tools. The research proposed a structured procedure to use existing digital technologies to support product-process design to analyse the workers behaviours and assess the perceived experience for industrial scopes. The paper defined a structured protocol analysis to objectify and measure the workers’ experience with the final aim to support the requirements definition in product-process design by using digital technologies. In particular, the model has been defined for the automotive sector. The paper contribution is the definition of the protocol analysis and the development of a mixed reality (MR) set- up to involve real users’ and to improve the digital models. Such a protocol has been applied to different industrial cases related to product and process design, developed in collaboration with CNH Industrial. The comparison with traditional design procedures highlighted the benefits of adopting virtual mock-up and digital simulation within a MR environment to shorten design time and improve the design overall quality
On the integrated design and development of a high speed cartesian robot
No full textNowadays, high speed cartesian robots are widely used as system integration of machine tools, commonly preferred to other robot architecture because of their high accuracy, reduced overall dimensions, and especially cost effectiveness and versatility. The increasing cycle speed and structural complexity of last generation machining processes imposes the design engineers to strongly improve the manipulators performances, and new requirements must absolutely be satisfied, for example the need to adopt proper safety systems and to implement different standard components to harmonically integrate with the existing machine tools. Since cartesian architecture robots are basically well developed industrial products, often already optimized during past design experiences, it becomes quite hard to find new solution and to further optimize the existing ones to realize superior performances and cost reduction. Then it becomes necessary to adopt a proper approach, on one hand able to follow the design engineers team intent during the fast evolving conceptual design stage and to help them to recognize, in a fast and effective way, the best solution to develop, on the other hand to optimize the design variant adopted. All the future users needs and design requirements must then absolutely be formalized, organized and correctly connected, in order to easily identify the task to perform, with the related priority, and their effects on all the different functionalities. In order to realize an effective robot configuration all the possible solution must be checked, as well as the choice of the motion apparatus type or the spatial placement of all the different components. Finally, once the decision of the detailed architecture took place, an intensive optimization activity becomes necessary. It is clear that a real performance improvement can be achieved only in the first design stage, where can be found design solution that can avoid future problems and simplify the structure. To really test the effectiveness of the choice adopted, a complete virtual prototype, strictly connected with the 3D CAD model, has been developed and all the possible solution verified, and compared, trough the numerical simulations, aimed to evaluate the robot different design variant final dynamical performance. The interaction between the CAD model and the virtual prototype imposed the use of third party spreadsheet, able to modify the main parameters and to easily process the simulation data results. In this paper the methodology and the solutions adopted for the design of a novel model of High Speed Cartesian Robot will be described
An Integrated Tool for the Virtual Commissioning of Flexible Robotic Cells
No full textIn the modern manufacturing industry, virtual models are normally used in machine and robot design for behavioral analysis and optimization. However, the literature reveals a notable gap in methodologies for simulating and conducting virtual commissioning multi-brand robotic systems with distributed control in a unified software environment before their physical installation. In this context, the present works leverages the open-source robotic libraries within the Robot Operating System (ROS) environment to extend the simulation capabilities of RoboDK software, aiming to support the validation and coordination of complex automated systems. This methodology is validated through procedures for conducting virtual commissioning of a robotic cell for an aircraft fuselage assembly whose process coordinator is defined in ROS. The obtained results demonstrate the effectiveness of the proposed simulation framework
- …
