Hasil untuk "Automation"

Jez Humble, D. Farley

N. Ivanova, Jeremy R. deWaard, P. Hebert

P. Tavares, C. M. Costa, Luís Rocha et al.

The optimization of the information flow from the initial design and through the several production stages plays a critical role in ensuring product quality while also reducing the manufacturing costs. As such, in this article we present a cooperative welding cell for structural steel fabrication that is capable of leveraging the Building Information Modeling (BIM) standards to automatically orchestrate the necessary tasks to be allocated to a human operator and a welding robot moving on a linear track. We propose a spatial augmented reality system that projects alignment information into the environment for helping the operator tack weld the beam attachments that will be later on seam welded by the industrial robot. This way we ensure maximum flexibility during the beam assembly stage while also improving the overall productivity and product quality since the operator no longer needs to rely on error prone measurement procedures and he receives his tasks through an immersive interface, relieving him from the burden of analyzing complex manufacturing design specifications. Moreover, no expert robotics knowledge is required to operate our welding cell because all the necessary information is extracted from the Industry Foundation Classes (IFC), namely the CAD models and welding sections, allowing our 3D beam perception systems to correct placement errors or beam bending, which coupled with our motion planning and welding pose optimization system ensures that the robot performs its tasks without collisions and as efficiently as possible while maximizing the welding quality.

Sobhana Mummaneni, Venkata Chaitanya Satya Ramaraju Mudunuri, Sri Veerabhadra Vikas Bommaganti et al.

A facial recognition system is a biometric security and surveillance system that can identify and monitor individuals in a crowded area. Manually monitoring a crowded environment is a difficult and error-prone task. Therefore, in such contexts, a model that automatically detects and recognises people's faces is needed to improve security. The automation of face recognition brings the benefit of a more efficient and accurate solution. This paper proposes an advanced model that has the ability to detect and recognise faces in dense crowds by using deep learning techniques. Where the input is live video, the process involves splitting the video into frames and each frame is fed into the model. The Multi-Task Cascaded Convolutional Neural Networks (MTCNN) algorithm is used for face detection. It accurately locates faces in frames and images and generates boundaries around the faces as output. The detected faces are then fed as input to a model, where they are compared with data from the database. If a face is recognised, the name of the recognised person is displayed in the boundary box of the frame, otherwise it is displayed that the person is unknown. FaceNet is used for face recognition tasks.

MA Changqing, LI Xuyang, LI Feng et al.

This study aims to accurately perceive the position and posture information of hydraulic supports in a disturbed environment. To address this, a precise perception method for the position and posture of hydraulic supports based on multi-sensor fusion was proposed. Firstly, nine-axis attitude sensors were deployed on four components of the hydraulic support, including top beam, shield beam, rear linkage, and base, to measure roll, pitch, and yaw angles using gyroscopes, accelerometers, and magnetometers. Then, the position and posture data was filtered using the Unscented Kalman Filter (UKF) algorithm and Improved Gradient Descent (IGD) algorithm (IGD-UKF algorithm), reducing interference from disturbance factors. Finally, an adaptive weighted fusion algorithm was employed to merge the filtered yaw and roll angle data of the top beam and base of hydraulic supports, eliminating data deviations caused by external vibrations, noise, and other factors. Perception experiments were conducted on the position and posture of top beam, shield beam, rear linkage, and base under various working conditions. The disturbances included the lowering and raising of top beam and base, as well as left-leaning, right-leaning, left-deviating and right-deviating of hydraulic supports. The study found that the data curves processed by the IGD-UKF algorithm exhibited smoother fluctuations, significantly suppressing oscillations and reducing amplitude. The yaw angle error of hydraulic supports ranged from 0.001 8° to 0.025 1°, with an average absolute error of 0.004 8°. The roll angle error ranged from 0.001 4° to 0.028 1°, with an average absolute error of 0.004 7°. The results indicate that the precise perception of the position and posture of hydraulic supports in a disturbed environment is achieved.

Ardian Kelmendi, George Pappas

The automotive industry increasingly relies on 3D modeling technologies to design and manufacture vehicle components with high precision. One critical challenge is optimizing the placement of latches that secure the dashboard side paneling, as these placements vary between models and must maintain minimal tolerance for movement to ensure durability. While generative artificial intelligence (AI) has advanced rapidly in generating text, images, and video, its application to creating accurate 3D CAD models remains limited. This paper proposes a novel framework that integrates a PointNet deep learning model with Python-based CAD automation to predict optimal clip placements and surface thickness for dashboard side panels. Unlike prior studies that focus on general-purpose CAD generation, this work specifically targets automotive interior components and demonstrates a practical method for automating part design. The approach involves generating placement data—potentially via generative AI—and importing it into the CAD environment to produce fully parameterized 3D models. Experimental results show that the prototype achieved a 75% success rate across six of eight test surfaces, indicating strong potential despite the limited sample size. This research highlights a clear pathway for applying generative AI to part design automation in the automotive sector and offers a foundation for scaling to broader design applications.

Wenyang Deng, Dongliang Xiao, Mingli Chen et al.

As distributed photovoltaic and shared energy storage systems expanded on the user side, developing an energy-sharing mechanism across different regions became crucial for fully utilizing local renewable energy resources and maximizing the system’s overall economic performance. This paper established a multi-regional energy operator (MREO) model considering shared energy storage, and a two-layer trading and optimization framework based on a master–slave game was developed. Initially, a trading system was devised to evaluate the interests of the power grid, MREO, and end-users. Next, an optimization model was formulated to capture the dynamic interactions between MREO decisions and user responses. The top-layer model was managed by MREO and focused on energy sharing among regions, which is used to set flexible electricity prices according to regional demand and optimize the use of shared energy storage. Meanwhile, the bottom-layer model addressed user demand response, allowing users to modify their energy consumption and select more advantageous trading areas based on information provided by the MREO. Simulation results confirmed that the proposed model accurately evaluated each party’s income, iteratively balanced their interests, and increased economic returns for both users and MREO. Additionally, the proposed approach supported greater local photovoltaic energy consumption, reduced grid load fluctuations, and fostered mutually beneficial outcomes for all stakeholders.

Andrew Prahl, L. M. Swol

D. Lyell, E. Coiera

Qian Chen, Borja García de Soto, B. Adey

Abstract Construction automation has shown the potential to increase construction productivity after years of technical development and experimenting in its field. Exactly how, and the possible benefits and challenges of construction automation, though is unclear and missing from current research efforts. In order to better understand the comprehensive potential of construction automation for increasing construction productivity and the associated possible ramifications, an objective and data-driven review of the use of automation technologies in construction was done. The review was accomplished by using text mining methods on publically available written documents, covering a wide range of relevant data including scientific publications and social media. The text mining software VOS Viewer and RapidMiner Studio were used to determine the most promising areas of research through the analysis of scientific publications, and the main areas of concern of industry through the analysis of text on social media, respectively. These research areas and concerns are summarized in this paper, and based on them suggestions for industry are made to help advance the uptake of automation in construction.

Scott A. Wright, Ainslie E. Schultz

Recent advancements in robotics, artificial intelligence, machine learning, and sensors now enable machines to automate activities that once seemed safe from disruption—including tasks that rely on higher-level thinking, learning, tacit judgment, emotion sensing, and even disease detection. Despite these advancements, the ethical issues of business automation and artificial intelligence—and who will be affected and how—are less understood. In this article, we clarify and assess the cultural and ethical implications of business automation for stakeholders ranging from laborers to nations. We define business automation and introduce a novel framework that integrates stakeholder theory and social contracts theory. By integrating these theoretical models, our framework identifies the ethical implications of business automation, highlights best practices, offers recommendations, and uncovers areas for future research. Our discussion invites firms, policymakers, and researchers to consider the ethical implications of business automation and artificial intelligence when approaching these burgeoning and potentially disruptive business practices.

T. Victor, E. Tivesten, P. Gustavsson et al.

Objective: The aim of this study was to understand how to secure driver supervision engagement and conflict intervention performance while using highly reliable (but not perfect) automation. Background: Securing driver engagement—by mitigating irony of automation (i.e., the better the automation, the less attention drivers will pay to traffic and the system, and the less capable they will be to resume control) and by communicating system limitations to avoid mental model misconceptions—is a major challenge in the human factors literature. Method: One hundred six drivers participated in three test-track experiments in which we studied driver intervention response to conflicts after driving highly reliable but supervised automation. After 30 min, a conflict occurred wherein the lead vehicle cut out of lane to reveal a conflict object in the form of either a stationary car or a garbage bag. Results: Supervision reminders effectively maintained drivers’ eyes on path and hands on wheel. However, neither these reminders nor explicit instructions on system limitations and supervision responsibilities prevented 28% (21/76) of drivers from crashing with their eyes on the conflict object (car or bag). Conclusion: The results uncover the important role of expectation mismatches, showing that a key component of driver engagement is cognitive (understanding the need for action), rather than purely visual (looking at the threat), or having hands on wheel. Application: Automation needs to be designed either so that it does not rely on the driver or so that the driver unmistakably understands that it is an assistance system that needs an active driver to lead and share control.

X. J. Yang, Vaibhav Unhelkar, Kevin Li et al.

Existing research assessing human operators' trust in automation and robots has primarily examined trust as a steady-state variable, with little emphasis on the evolution of trust over time. With the goal of addressing this research gap, we present a study exploring the dynamic nature of trust. We defined trust of entirety as a measure that accounts for trust across a human's entire interactive experience with automation, and first identified alternatives to quantify it using real-time measurements of trust. Second, we provided a novel model that attempts to explain how trust of entirety evolves as a user interacts repeatedly with automation. Lastly, we investigated the effects of automation transparency on momentary changes of trust. Our results indicated that trust of entirety is better quantified by the average measure of “area under the trust curve” than the traditional post-experiment trust measure. In addition, we found that trust of entirety evolves and eventually stabilizes as an operator repeatedly interacts with a technology. Finally, we observed that a higher level of automation transparency may mitigate the “cry wolf’ effect - wherein human operators begin to reject an automated system due to repeated false alarms.

Lei Chen, Yunchen Yu, Jie Luo et al.

The vehicle dynamics model has multiple degrees of freedom, with strong nonlinear characteristics, so it is difficult to quickly obtain the accurate target oil pressure of an electronically assisted brake system based on the model. This paper proposes a target oil pressure recognition algorithm based on the T-S fuzzy neural network model. Firstly, the braking conditions classification algorithm is built according to the sampled braking intention data. The data are divided into the emergency braking condition data and the general braking condition data by the braking conditions classification algorithm. Secondly, the recognition model is trained respectively by the different braking condition data sets. In the training process, the fuzzy C-means clustering algorithm is used to identify the antecedent parameters of the model, and the learning rate cosine attenuation strategy is applied to optimize the parameter learning process. Finally, a correction method of target oil pressure based on slip ratio is proposed, and the target oil pressure derived following control methods based on traditional PID and fuzzy PID are compared through experiments. The results show that the mean square error of oil pressure following control based on fuzzy PID is smaller, which proves that the proposed method is able to precisely control braking force.

T. Louw, N. Merat

This driving simulator study, conducted as part of the EC-funded AdaptIVe project, assessed drivers’ visual attention distribution during automation and on approach to a critical event, and examined whether such attention changes following repeated exposure to an impending collision. Measures of drivers’ horizontal and vertical gaze dispersion during both conventional and automated (SAE Level 2) driving were compared on approach to such critical events. Using a between-participant design, 60 drivers (15 in each group) experienced automation with one of four screen manipulations: (1) no manipulation, (2) manipulation by light fog, (3) manipulation by heavy fog, and (4) manipulation by heavy fog with a secondary task, which were used to induce varying levels of engagement with the driving task. Results showed that, during automation, drivers’ horizontal gaze was generally more dispersed than that observed during manual driving. Drivers clearly looked around more when their view of the driving scene was completely blocked by an opaque screen in the heavy fog condition. By contrast, horizontal gaze dispersion was (unsurprisingly) more concentrated when drivers performed a visual secondary task, which was overlaid on the opaque screen. However, once the manipulations ceased and an uncertainty alert captured drivers’ attention towards an impending incident, a similar gaze pattern was found for all drivers, with no carry-over effects observed after the screen manipulations. Results showed that drivers’ understanding of the automated system increased as time progressed, and that scenarios that encourage driver gaze towards the road centre are more likely to increase situation awareness during high levels of automation.

Sebastian Hergeth, Lutz Lorenz, J. Krems

Toby Jia-Jun Li, A. Azaria, B. Myers

Vittoria Garzelli

Annette M. O’Connor, G. Tsafnat, James Thomas et al.

BackgroundAlthough many aspects of systematic reviews use computational tools, systematic reviewers have been reluctant to adopt machine learning tools.DiscussionWe discuss that the potential reason for the slow adoption of machine learning tools into systematic reviews is multifactorial. We focus on the current absence of trust in automation and set-up challenges as major barriers to adoption. It is important that reviews produced using automation tools are considered non-inferior or superior to current practice. However, this standard will likely not be sufficient to lead to widespread adoption. As with many technologies, it is important that reviewers see “others” in the review community using automation tools. Adoption will also be slow if the automation tools are not compatible with workflows and tasks currently used to produce reviews. Many automation tools being developed for systematic reviews mimic classification problems. Therefore, the evidence that these automation tools are non-inferior or superior can be presented using methods similar to diagnostic test evaluations, i.e., precision and recall compared to a human reviewer. However, the assessment of automation tools does present unique challenges for investigators and systematic reviewers, including the need to clarify which metrics are of interest to the systematic review community and the unique documentation challenges for reproducible software experiments.ConclusionWe discuss adoption barriers with the goal of providing tool developers with guidance as to how to design and report such evaluations and for end users to assess their validity. Further, we discuss approaches to formatting and announcing publicly available datasets suitable for assessment of automation technologies and tools. Making these resources available will increase trust that tools are non-inferior or superior to current practice. Finally, we identify that, even with evidence that automation tools are non-inferior or superior to current practice, substantial set-up challenges remain for main stream integration of automation into the systematic review process.

B. Vermeulen, Jan Kesselhut, A. Pyka et al.

We study the projected impact of automation on employment in the forthcoming decade, both at the macro-level and in actual (types of) sectors. Hereto, we unite an evolutionary economic model of multisectoral structural change with labor economic theory. We thus get a comprehensive framework of how displacement of labor in sectors of application is compensated by intra- and intersectoral countervailing effects and notably mopped up by newly created, labor-intensive sectors. We use several reputable datasets with expert projections on employment in occupations affected by automation (and notably by the introduction of robotics and AI) to pinpoint which and how sectors and occupations face employment shifts. This reveals how potential job loss due to automation in “applying” sectors is counterbalanced by job creation in “making” sectors as well in complementary and quaternary, spillover sectors. Finally, we study several macro-level scenarios on employment and find that mankind is facing “the usual structural change” rather than the “end of work”. We provide recommendations on policy instruments that enhance the dynamic efficiency of structural change.