Hasil untuk "Machine design and drawing"

Bougault Pierrick, Haddad Raphael El, Ma Liang

Artificial Intelligence (AI) and Machine Learning (ML) are foundations in new manufacturing paradigms, yet their application in the aircraft industry remains limited, as this industry’s expertise does not traditionally cover these technologies. Additionally, due to its specific features, the aircraft industry presents unique challenges, for instance, with data scarcity. To date, no systematic review has considered these features to enable stakeholders in this sector to undergo AI/ML transformation successfully. This study aims to analyze the state of the art by providing a PRISMA systematic literature review of 135 articles, focusing on the contexts, models, and methods employed in the development of AI/ML solutions. The authors propose a framework to summarize the findings on the development, applications, benefits, and challenges of AI/ML in the aircraft manufacturing industry. In addition, further research opportunities are identified through a comparison of current research applications, theoretical concepts of Industry 5.0, and cutting-edge technologies, such as Federated Learning, Transfer Learning, the use of Large Language Models (LLMs), the lack of supply chain investigation, and the integration of human factors, which are absent in major reviewed articles. This study contributes to the field by meticulously gathering methodologies and approaches that address and integrate the specificities of AI/ML use and integration in this high-value-added industry. It bridges the gap between cutting-edge research and practical industry needs, delivering actionable insights to drive innovation and guide strategic decision-making.

Moldakhanov Bekbolat, Asangaliev Yelibek, Kim Alina et al.

This article examines the effect of the mixing degree of feed mixture components on its homogeneity, mixing quality, and the uniform distribution of microcomponents and vitamins within the volume of the mixture. Additionally, the study evaluates the suitability of the mixture for the molding and pressing processes involved in the production of salt lick bricks (SLB). The research is conducted using a novel energy-efficient combined mixer. The study presents an analysis of the impact of mixing intensity on the quality of SLB under prolonged storage in various environmental conditions. The kinetics of the mixing process for salt premix components is investigated. A mechanical-mathematical model describing the mixing behavior of components in a combined mixer is developed. Based on experimental data, correlations are established between the physical characteristics of the final salt lick premixes – including shape, composition, and hardness – and the homogeneity of the initial mixture. Furthermore, the study assesses the stability of SLB premixes under adverse environmental conditions, including exposure to humidity and temperature fluctuations. A methodology for determining the optimal rotor rotation frequency in the combined mixer is developed, and its influence on mixture homogeneity is quantified.

Basheer Wasef Shaheen, Ahmed Jaber

Electric scooters and other micromobility innovations are becoming standard fare in urban transportation networks. Yet there are several obstacles that must be overcome, including concerns about users’ satisfaction and safety. This study aimed primarily at developing a user-centered methodological framework that combined different user-centered engineering tools such as voice of customers analysis, needs–metrics mapping, Pugh’s matrix and morphological design, strategic analysis approaches such as SWOT and PESTEL, and, a key innovation, the smart terrain-adaptive power management system (STAPMS), an AI-based feature that dynamically adjusts power output and regenerative braking based on Budapest’s varied topography and road conditions to improve energy efficiency and ride comfort. This innovative framework offers insights into redesign options aimed at enhancing customer satisfaction, product quality, and business growth. The proposed framework was validated on Lime electric scooters, particularly the S2 generation type. Three design concepts were generated and evaluated through a systematic approach to provide an optimal balance between users’ needs, technical performance, and strategic feasibility. The proposed user-centered framework shows significant potential to improve users’ satisfaction, enhanced usability, extended range, and increased market competitiveness, validating its viability for micromobility innovative solutions. The findings also demonstrate the necessity for systematic frameworks that link user experience with engineering design and can be generalized to other micromobility products.

Junfeng Hu, Chao Liu

Abstract This paper presents a bio-inspired soft climbing robot designed to overcome the limitations of conventional climbing robots in unstructured, irregular, or deformable environments. Mimicking the arboreal locomotion of sloths, the robot employs an embracing-based anchoring mechanism driven by a tensegrity-structured spiral system, capable of generating gripping forces exceeding 20 N through topological interlocking. Coupled with a telescopic actuation module utilizing tunable tower springs, the robot achieves adaptive body deformation and bidirectional locomotion. Experimental results demonstrate stable vertical climbing, passive adaptation to curved and variable-diameter surfaces, active gap-crossing via elastic energy release, and reliable performance on deformable substrates. This work establishes a new paradigm for robotic locomotion in complex 3D environments by integrating bio-inspired topological constraints with elastic actuation.

Till Temmen, Jasper Debougnoux, Li Li et al.

The development of AI-driven automated driving functions requires vast amounts of diverse, high-quality data to ensure road safety and reliability. However, both the manual collection of real-world data and creation of 3D environments are costly, time-consuming, and hard to scale. Most automatic environment generation methods still rely heavily on manual effort, and only a few are tailored for Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) training and validation. We propose an automated generative framework that learns ground-truth features to reconstruct 3D environments from a road definition and two simple parameters for country and area type. Environment generation is structured into three modules—map-based data generation, semantic city generation, and final detailing. The overall framework is validated by training a perception network on a mixed set of real and synthetic data, validating it solely on real data, and comparing performance to assess the practical value of the environments we generated. By constructing a Pareto front over combinations of training set sizes and real-to-synthetic data ratios, we show that our synthetic data can replace up to <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>85</mn><mo>%</mo></mrow></semantics></math></inline-formula> of real data without significant quality degradation. Our results demonstrate how multi-layered environment generation frameworks enable flexible and scalable data generation for perception tasks while incorporating ground-truth 3D environment data. This reduces reliance on costly field data and supports automated rapid scenario exploration for finding safety-critical edge cases.

Adam Skokan

This paper evaluates Advanced Driver Assistance Systems (ADASs) in test scenarios derived from naturalistic driving and crash data, mapped to ISO 26262, ISO/PAS 21448 (SOTIF), and ISO 34502. From eight high-risk scenarios, it is validated for left turns across oncoming traffic on a proving ground using a Škoda Superb iV against a soft Global Vehicle Target. ODD and spatiotemporal thresholds are parameterized and speed/acceleration profiles from GNSS/IMU data are analyzed. AEB and FCW performance varies across nominally identical runs, driven by human-in-the-loop variability and target detectability. In successful interventions, peak deceleration reached −0.64 g, meeting UNECE R152 criteria; in other runs, late detection narrowed TTC below intervention thresholds, leading to contact. Limitations in current protocols are identified and argue for scenario catalogs with realistic context (weather, surface, masking) and latency-aware metrics. The results motivate extending validation beyond standard tracks toward mixed methods linking simulation, scenario databases, and instrumented field trials.

Eugen Valentin Butilă, Răzvan Gabriel Boboc

The automotive sector plays an essential role in the Romanian economy, making a significant contribution to industrial production and employment. This study conducts a comprehensive bibliometric analysis of scholarly publishing in the Romanian automotive sector. By analyzing publication trends, citation patterns, and collaboration networks, the study maps the evolution of research in this field and highlights key contributions and future directions. The findings reveal a significant increase in research output over the past two decades, with a focus on emerging fields such as artificial intelligence, electric and autonomous vehicles, and sustainable mobility solutions. The analysis also identifies leading researchers and institutions and explores collaboration networks between Romanian and international actors. These insights provide valuable benchmarks for assessing Romania’s position in the global automotive research arena and inform strategies for future research efforts.

Miroslav Penchev, Kent C. Johnson, Arun S. K. Raju et al.

This study presents a well-to-wheel life-cycle assessment (WTW-LCA) comparing battery-electric heavy-duty trucks (BEVs) with conventional diesel trucks, utilizing real-world fleet data from Southern California’s Volvo LIGHTS project. Class 7 and Class 8 vehicles were analyzed under ISO 14040/14044 standards, combining measured diesel emissions from portable emissions measurement systems (PEMSs) with BEV energy use derived from telematics and charging records. Upstream (“well-to-tank”) emissions were estimated using USLCI datasets and the 2020 Southern California Edison (SCE) power mix, with an additional scenario for BEVs powered by on-site solar energy. The analysis combines measured real-world energy consumption data from deployed battery electric trucks with on-road emission measurements from conventional diesel trucks collected by the UCR team. Environmental impacts were characterized using TRACI 2.1 across climate, air quality, toxicity, and fossil fuel depletion impact categories. The results show that BEVs reduce total WTW CO<sub data-eusoft-scrollable-element="1">2</sub>-equivalent emissions by approximately 75% compared to diesel. At the same time, criteria pollutants (NO<sub data-eusoft-scrollable-element="1">x</sub>, VOCs, SO<sub data-eusoft-scrollable-element="1">x</sub>, PM<sub data-eusoft-scrollable-element="1">2.5</sub>) decline sharply, reflecting the shift in impacts from vehicle exhaust to upstream electricity generation. Comparative analyses indicate BEV impacts range between 8% and 26% of diesel levels across most environmental indicators, with near-zero ozone-depletion effects. The main residual hotspot appears in the human-health cancer category (~35–38%), linked to upstream energy and materials, highlighting the continued need for grid decarbonization. The analysis focuses on operational WTW impacts, excluding vehicle manufacturing, battery production, and end-of-life phases. This use-phase emphasis provides a conservative yet practical basis for short-term fleet transition strategies. By integrating empirical performance data with life-cycle modeling, the study offers actionable insights to guide electrification policies and optimize upstream interventions for sustainable freight transport. These findings provide a quantitative decision-support basis for fleet operators and regulators planning near-term heavy-duty truck electrification in regions with similar grid mixes, and can serve as an empirical building block for future cradle-to-grave and dynamic LCA studies that extend beyond the operational well-to-wheels scope adopted here.

Afra Kilic, Kim Batselier

Tensor Network (TN) Kernel Machines speed up model learning by representing parameters as low-rank TNs, reducing computation and memory use. However, most TN-based Kernel methods are deterministic and ignore parameter uncertainty. Further, they require manual tuning of model complexity hyperparameters like tensor rank and feature dimensions, often through trial-and-error or computationally costly methods like cross-validation. We propose Bayesian Tensor Network Kernel Machines, a fully probabilistic framework that uses sparsity-inducing hierarchical priors on TN factors to automatically infer model complexity. This enables automatic inference of tensor rank and feature dimensions, while also identifying the most relevant features for prediction, thereby enhancing model interpretability. All the model parameters and hyperparameters are treated as latent variables with corresponding priors. Given the Bayesian approach and latent variable dependencies, we apply a mean-field variational inference to approximate their posteriors. We show that applying a mean-field approximation to TN factors yields a Bayesian ALS algorithm with the same computational complexity as its deterministic counterpart, enabling uncertainty quantification at no extra computational cost. Experiments on synthetic and real-world datasets demonstrate the superior performance of our model in prediction accuracy, uncertainty quantification, interpretability, and scalability.

Sarah Segel, Helena Graf, Edward Bergman et al.

Hyperparameter optimization (HPO), as a central paradigm of AutoML, is crucial for leveraging the full potential of machine learning (ML) models; yet its complexity poses challenges in understanding and debugging the optimization process. We present DeepCAVE, a tool for interactive visualization and analysis, providing insights into HPO. Through an interactive dashboard, researchers, data scientists, and ML engineers can explore various aspects of the HPO process and identify issues, untouched potentials, and new insights about the ML model being tuned. By empowering users with actionable insights, DeepCAVE contributes to the interpretability of HPO and ML on a design level and aims to foster the development of more robust and efficient methodologies in the future.

Alex Gower

Physical systems that naturally perform energy descent offer a direct route to accelerating machine learning. Oscillator Ising Machines (OIMs) exemplify this idea: their GHz-frequency dynamics mirror both the optimization of energy-based models (EBMs) and gradient descent on loss landscapes, while intrinsic noise corresponds to Langevin dynamics - supporting sampling as well as optimization. Equilibrium Propagation (EP) unifies these processes into descent on a single total energy landscape, enabling local learning rules without global backpropagation. We show that EP on OIMs achieves competitive accuracy ($\sim 97.2 \pm 0.1 \%$ on MNIST, $\sim 88.0 \pm 0.1 \%$ on Fashion-MNIST), while maintaining robustness under realistic hardware constraints such as parameter quantization and phase noise. These results establish OIMs as a fast, energy-efficient substrate for neuromorphic learning, and suggest that EBMs - often bottlenecked by conventional processors - may find practical realization on physical hardware whose dynamics directly perform their optimization.

Till Korten, Vladimir Rybnikov, Mathias Vogt et al.

Electron beam accelerators are essential in many scientific and technological fields. Their operation relies heavily on the stability and precision of the electron beam. Traditional diagnostic techniques encounter difficulties in addressing the complex and dynamic nature of electron beams. Particularly in the context of free-electron lasers (FELs), it is fundamentally impossible to measure the lasing-on and lasingoff electron power profiles for a single electron bunch. This is a crucial hurdle in the exact reconstruction of the photon pulse profile. To overcome this hurdle, we developed a machine learning model that predicts the temporal power profile of the electron bunch in the lasing-off regime using machine parameters that can be obtained when lasing is on. The model was statistically validated and showed superior predictions compared to the state-of-the-art batch calibrations. The work we present here is a critical element for a virtual pulse reconstruction diagnostic (VPRD) tool designed to reconstruct the power profile of individual photon pulses without requiring repeated measurements in the lasing-off regime. This promises to significantly enhance the diagnostic capabilities in FELs at large.

Taichi Yamada, Hiroyuki Abe, Katsuji Ogane et al.

Abstract This paper introduces the development of test methods for capability evaluation of Unmanned Aircraft Systems (UASs) in Global Navigation Satellite System (GNSS)-denied environments. The purpose of this development is to facilitate growth in the UAS industry. We discuss the test method’s approach and what UAS’s capability evaluation is essential for UAS use in GNSS-denied environments. As a result, we decided to adopt an approach that a test method evaluates a capability to perform a single simple task. In addition, these test methods should be along with the requirement of UAS manufacturers and users. Thus, first, we develop test methods to assess mobility in narrow spaces with obstacles shielding GNSS radio. We repeatedly have demonstrations and discussions of our test method with UAS manufacturers and users from the early stage of this development and collect their opinions for improvement to proceed with the development while building consensus with them. In addition, we evaluate several UASs by the test methods and examine whether our test method makes it possible to show the difference in UAS capability. This paper describes the approach of the test methods, the test methods for mobility in a GNSS-denied environment, the demonstrations and discussions with UAS manufacturers and users, and the results of the UAS’s performance evaluation by the test methods.

Dimitrios Koulocheris, Clio Vossou

In the literature the value of the driver’s head acceleration has been widely used as an objective function for the modification of the suspension and/or the seat characteristics in order to optimize the ride comfort of a vehicle. For these optimization procedures various lumped parameter Vehicle–Seat–Human models are proposed. In the present paper a Quarter Car model is integrated with three Seat–Human models with different levels of detail. The level of detail corresponds to the number of degrees of freedom used to describe the Seat–Human system. Firstly, the performance of the Quarter Car model, used as a basis, is analyzed in six excitations with different characteristics. Then, the performance of the three lumped parameter Vehicle–Seat–Human models are monitored in the same excitations. The results indicated that in the case of single disturbance excitations the Quarter Car model provided 50–75% higher values of acceleration compared with the eight degrees of freedom model. As far as the periodic excitation is concerned, the Vehicle–Seat–Human models provided values of acceleration up to eight times those of the Quarter Car model. On the other hand, in stochastic excitations the Vehicle–Seat–Human model with three degrees of freedom produced the closest results to the Quarter Car model followed by the eight degrees of freedom model. Finally, with respect to the computational efficiency it was found that an increase in the degrees of freedom of the Vehicle–Seat–Human model by one caused an increase in the CPU time from 2.1 to 2.6%, while increasing the number of the degrees of freedom by five increased the CPU time from 7.4 to 11.5% depending on the excitation.

Abdennour Boulesnane

Evolutionary Computation (EC) has emerged as a powerful field of Artificial Intelligence, inspired by nature's mechanisms of gradual development. However, EC approaches often face challenges such as stagnation, diversity loss, computational complexity, population initialization, and premature convergence. To overcome these limitations, researchers have integrated learning algorithms with evolutionary techniques. This integration harnesses the valuable data generated by EC algorithms during iterative searches, providing insights into the search space and population dynamics. Similarly, the relationship between evolutionary algorithms and Machine Learning (ML) is reciprocal, as EC methods offer exceptional opportunities for optimizing complex ML tasks characterized by noisy, inaccurate, and dynamic objective functions. These hybrid techniques, known as Evolutionary Machine Learning (EML), have been applied at various stages of the ML process. EC techniques play a vital role in tasks such as data balancing, feature selection, and model training optimization. Moreover, ML tasks often require dynamic optimization, for which Evolutionary Dynamic Optimization (EDO) is valuable. This paper presents the first comprehensive exploration of reciprocal integration between EDO and ML. The study aims to stimulate interest in the evolutionary learning community and inspire innovative contributions in this domain.

Aruna Jayasena, Prabhat Mishra

Increasing design complexity and reduced time-to-market have motivated manufacturers to outsource some parts of the System-on-Chip (SoC) design flow to third-party vendors. This provides an opportunity for attackers to introduce hardware Trojans by constructing stealthy triggers consisting of rare events (e.g., rare signals, states, and transitions). There are promising test generation-based hardware Trojan detection techniques that rely on the activation of rare events. In this paper, we investigate rareness reduction as a design-for-trust solution to make it harder for an adversary to hide Trojans (easier for Trojan detection). Specifically, we analyze different avenues to reduce the potential rare trigger cases, including design diversity and area optimization. While there is a good understanding of the relationship between area, power, energy, and performance, this research provides a better insight into the dependency between area and security. Our experimental evaluation demonstrates that area reduction leads to a reduction in rareness. It also reveals that reducing rareness leads to faster Trojan detection as well as improved coverage by Trojan detection methods.

Mohsin Yousufi, Charlotte Alexander, Nassim Parvin

This paper brings attention to epistemic injustice, an issue that has not received much attention in the design of technology and policy. Epistemic injustices occur when individuals are treated unfairly or harmed specifically in relation to their role as knowers or possessors of knowledge. Drawing on the case of making heat complaints in New York City, this paper illustrates how both technological and policy interventions that address epistemic injustice can fail or even exacerbate the situations for certain social groups, and individuals within them. In bringing this case to the workshop, this paper hopes to provide another generative and critical dimension that can be utilised to create better technologies and policies, especially when they deal with diverse and broad range of social groups

Tobias Eichenlaub, Paul Heckelmann, Stephan Rinderknecht

Driving style and external factors such as traffic density have a significant influence on the vehicle energy demand especially in city driving. A longitudinal control approach for intelligent, connected vehicles in urban areas is proposed in this article to improve the efficiency of automated driving. The control approach incorporates information from Vehicle-2-Everything communication to anticipate the behavior of leading vehicles and to adapt the longitudinal control of the vehicle accordingly. A supervised learning approach is derived to train a neural prediction model based on a recurrent neural network for the speed trajectories of the ego and leading vehicles. For the development, analysis and evaluation of the proposed control approach, a co-simulation environment is presented that combines a generic vehicle model with a microscopic traffic simulation. This allows for the simulation of vehicles with different powertrains in complex urban traffic environment. The investigation shows that using V2X information improves the prediction of vehicle speeds significantly. The control approach can make use of this prediction to achieve a more anticipatory driving in urban areas which can reduce the energy consumption compared to a conventional Adaptive Cruise Control approach.

Umberto Previti, Antonio Galvagno, Giacomo Risitano et al.

This work aimed to develop an automatic new methodology based on establishing if a mechanical component, designed for a conventional propulsion system, is also suitable for hybrid electric propulsion. Change in propulsion system leads to different power delivery and vehicle dynamics, which will be reflected in different load conditions acting on the mechanical components. It has been shown that a workflow based on numerical simulations and experimental tests represents a valid approach for the evaluation of the cumulative fatigue damage of a mechanical component. In this work, the front half-shaft of a road car was analyzed. Starting from the acquisition of a speed profile and the definition of a reference vehicle, in terms of geometry and transmission, a numerical model, based on longitudinal vehicle dynamics, was developed for both conventional and hybrid electric transmission. After the validation of the model, the cumulative fatigue damage of the front half-shaft was evaluated. The new design methodology is agile and light; it has been dubbed “Smart Design”. The results show that changing propulsion led to greater fatigue damage, reducing the fatigue life component by 90%. Hence, it is necessary to redesign the mechanical component to make it also suitable for hybrid electric propulsion.

Felix Deufel, Purav Jhaveri, Marius Harter et al.

In order to improve the efficiency of electrified vehicle drives, various predictive energy management strategies (driving strategies) have been developed. This article presents the extension of a generic prediction approach already proposed in a previous paper, which allows a robust forecasting of all traction torque-relevant variables for such strategies. The extension primarily includes the proper utilization of map data in the case of an a priori known route. Approaches from Artificial Intelligence (AI) have proven to be effective for such proposals. With regard to this, Recurrent Neural Networks (RNN) are to be preferred over Feed-Forward Neural Networks (FNN). First, preprocessing is described in detail including a wide overview of both calculating the relevant quantities from global navigation satellite system (GNSS) data in several steps and matching these with data from the chosen map provider. Next, an RNN including Long Short-Term Memory (LSTM) cells in an Encoder–Decoder configuration and a regular FNN are trained and applied. The models are used to forecast real driving profiles over different time horizons, both including and excluding map data in the model. Afterwards, a comparison is presented, including a quantitative and a qualitative analysis. The accuracy of the predictions is therefore assessed using Root Mean Square Error (RMSE) computations and analyses in the time domain. The results show a significant improvement in velocity prediction with LSTMs including map data.