This study proposes a generalized Risk-Based Approach (RBA) methodology by adapting the development assurance processes used in aviation, such as ARP4754B and SORA, for broader industrial applications beyond safety-critical domains. The proposed method enables the optimization of resource allocation under high uncertainty by adjusting the rigor of response according to the impact level of each business risk. An application case in the sports industry demonstrates the method's utility in optimizing the quality standards of badminton rackets. As a result, 7 critical risks were identified for increased resource allocation, while 14 lower-impact risks were rationally simplified, reducing resource usage without compromising acceptable risk levels. The approach proved effective in achieving internal consensus, enabling simultaneous reinforcement and streamlining of processes. This research contributes to both academia and practice by offering a structured, scalable RBA framework applicable across domains, especially in innovation-driven product development under constraints. Future work will refine robustness quantification and explore applicability to other sectors such as services and infrastructure.

Powder compression molding is widely utilized as a food processing technology, and ensuring product strength is an essential issue for improving quality. The Drucker–Prager Cap (DPC) model is known as an analytical model for powder compression and can represent deformation behavior during the compression process. In this study, material properties (DPC parameters) of four types of food powders were derived using a tablet data acquisition device capable of simultaneously measuring compressive loads in both axial and radial directions. Furthermore, the DPC parameters were input into the CAE software Ansys, and compression analysis was conducted on a cuboid-shaped model. The strain characteristics obtained from the analysis were then compared with experimental data from friability test. The findings indicated that the higher the proportion of elastic strain to total strain at maximum compression, the lower the abrasion level tended to be. This study clarified the importance of the elastic properties of powder layers in determining product strength during the compression molding of food powders.

NV (Noise and Vibration) performance is one of the key product qualities of a mechanical system. Isolating all resonant frequencies from the frequency band where excitation forces are high is an important strategy for improving NV performance. However, a challenge in designing NV performance is that the resonant frequency is determined by the contributions of all the subsystems that make up the mechanical system. Complex mechanical systems such as automobiles are concurrently developed, in which the supplier subsystem is mounted onto the OEM (Original Equipment Manufacturer) subsystem. As a result, it is not easy to share intellectual property, such as shape information, between companies. For this reason, NV performance, including the assignment of resonant frequencies, is often evaluated in the final stage of product development. However, if NV performance targets are not met at this stage, extensive redevelopment will be required. Therefore, it is necessary to be able to design the NV performance of a whole structure from the supplier's viewpoint at the upstream product development stage. In this paper, we proposed a method to solve the underdetermined inverse problem of appropriately allocating the multiple resonant frequencies of a whole structure, which occurs due to a large contribution from the supplier subsystem, by modifying the structure of only the supplier subsystem. This method is composed of the kCA (kernel Compliance Analysis) and the CMCM (Cross-Model Cross-Mode) method and is realized by using the compliance-FRF matrix of the OEM subsystem provided by the OEM to the supplier in the upstream design stage. Finally, numerical verification of the proposed method was demonstrated.

With the further expansion of the Shinkansen network to northern region, sufficient low-temperature fluidity is required for axle-box oil of Shinkansen trains. Axle-box oil with excellent low-temperature performance has been developed in the past, however, it has not been put into practical use because of the maintenance problem caused by reddening when it is exposed to ultraviolet rays. Therefore, we have newly developed axle-box oils that are less prone to reddening and exhibit improved low-temperature performance compared with the current oil. The properties and performance of the prototype oils were evaluated in laboratory tests. The prototype oils exhibited improved low-temperature fluidity and did not redden under ultraviolet irradiation. The thermal stability and lubricating performance of the prototype oils was no less than those of the current oil. In addition, the durability tests using actual-size axle-box bearings were conducted for the more promising candidate of the prototypes. According to the test results, the durability of the prototype oil for 800,000 km use was confirmed because no abnormality was found from the results of the inspection of oils and bearings in between and after the test.

A novel force distribution adjustment mechanism for force sensors is proposed to implement variable admittance control in robotic systems. While admittance control is widely utilized in physical human–robot interactions (pHRI), existing approaches face significant limitations: software-based variable admittance control is constrained by the bandwidth of the robot controller, which can potentially cause instability in high-rigidity environments. Meanwhile, hardware solutions that utilize low-compliance mechanisms lack adaptability to varying operational conditions. This paper developed a novel force sensor-based mechanism that enables variable admittance control through mechanical adjustment of the force distribution between the detection and non-detection interfaces of a force sensor. The proposed mechanism, which is approximately 50% smaller than the force sensor, employs distributed load theory to adjust the gain by varying the contact surface interfaces between the detection and non-detection, enabling stable gain adjustment even in high-rigidity structures. In contrast to conventional beam-based approaches where the range of gain adjustment decreases with increased rigidity, the proposed method maintains variable characteristics regardless of the structural rigidity. Experimental validation through experiments for evaluating the gain adjustment characteristics and spiral groove tracing with Peg-in-Hole tasks experiments demonstrated the gain adjustment capability of the proposed mechanism and verified its implementation of variable admittance control through a simple hardware installation.

This paper presents the development and evaluation of a Mechanical Positioned Guided Vehicle (MPGV), a sensorless and mechanical-control mobile robot designed to achieve high-precision positioning of less than ±5 mm. AGVs and AMRs have traditionally required high-tech sensors and control systems for accurate positioning, however these technologies increase costs and require specialized knowledge for implementation and maintenance. The MPGV utilizes mechanical components and mechanisms to achieve accurate positioning without relying on electronic control and sensors. It features two driving modes: a normal mode, which uses common guidance methods like magnetic tapes, and a guide-rail mode, which mechanically ensures precise positioning. In guide-rail mode, the MPGV is constrained by a guide rail, allowing for straightforward, high-precision positioning using mechanical stops and a wheel lift-off mechanism to decouple power. The paper describes the design and functional principle of key mechanical systems, including transmission and wheel lift-off mechanisms, and conducts experiments to verify the MPGV’s functionality and precision. Experimental results show that the MPGV can switch between driving modes and achieve high-precision positioning within ±5 mm, confirming the feasibility of using mechanical solutions for industrial positioning applications without costly electronic control systems.

In the present study, we investigated the influence of the magnetization angle of cubic magnetic particles on the internal structure of aggregates by means of Monte Carlo and Brownian dynamics simulations. Monte Carlo method was used to examine the internal structure of aggregates at equilibrium, while Brownian dynamics method was employed to analyze the time evolution of the internal structure during the approach to equilibrium. Monte Carlo simulations showed that face-oriented magnetic moments promote the formation of chain-like clusters, whereas edge- and diagonal-oriented moments lead to closely-packed clusters. Brownian dynamics simulations revealed that the growth rate of aggregates depends on the magnetization direction. As the magnetization angle approached the face direction from the diagonal direction, the internal structure of the aggregates tended to transform from closely-packed clusters to chain-like and ring-like clusters. In particular, a distinct structural transition was observed when the magnetization angle was in the range of approximately 15° to 20° from the diagonal direction. In contrast, changing the magnetization angle from the diagonal direction to the edge direction had little effect on the internal structure of the aggregates, and closely-packed clusters were still observed. These findings suggested that a magnetization direction tilted by approximately 12° from the diagonal direction, as experimentally predicted, was close to the critical angle where a structural transition occurred. As a result, although closely-packed clusters remained the dominant structure, ring-like and chain-like clusters also formed simultaneously within the system.

In this study, we investigate the accuracy of four-node tetrahedral elements with drilling and strain degrees of freedom (GNTet4) for nearly incompressible materials. GNTet4 is a one of the generalized finite element method approximation. Three types of GNTet4 are introduced by applying approximation polynomials of the Taylor expansion. First, we reveal the extra degrees of freedom constraints without the zero-energy modes of each GNTet4 through eigenvalue analysis. Next, the displacement accuracy of each GNTet4 is evaluated by analyzing a cantilever beam. The results shows that GNTet4, which approximates the third-order Taylor expansion (GNTet4-3rd), is almost identical to a 20-node tetrahedral element. Furthermore, a plane-strain compression test is performed to investigate the analytical accuracy of each GNTet4 for nearly incompressible materials. To analyze nearly incompressible materials with high accuracy, selective reduced integration (SRI) and third-order Taylor expansion must be applied to GNTet4 (GNTet4-3rd with SRI). Incidentally, the analysis accuracy of GNTet4-3rd with SRI is the same as or higher than that of the 10-node tetrahedral element (FEM-Tet10). Furthermore, the GNT4-3rd with SRI, which possess the nodal point at the element vertex only, suppressed the nodal oscillation. Thus, in nearly incompressible materials, to obtain the same analysis accuracy as that of FEM-Tet10, the four-node tetrahedral element must apply the displacement function of the fourth order and the SRI.

With new services provided by railway companies and changes in lifestyles, passengers doing tasks while traveling on Shinkansen and other trains are increasing. However, the conventional method of providing information to passengers is through voice announcements and LCD monitors in the train, and when working remotely while participating in an online meeting in a train, vision and hearing are used at the same time. Passengers performing such tasks must temporarily stop their visual or auditory tasks in order to obtain such information. Therefore, in this research, the purpose is to obtain information sent to passengers within the railway vehicle without interrupting visual tasks such as using a laptop, which passengers are performing inside the railway vehicle by utilizing the characteristics of human visual perception, especially the characteristics of the effective field of view. Subjective evaluation, eye movement, task performance, and other indicators were measured for 15 experiment participants when information was presented using the HMI proposed in this study. From the analysis results of experimental conditions, subjective evaluation values, and behavioral evaluation values, we found that experiment participants recognize a pictogram with little psychological effect, and 61% of them do not have a movement of gaze point.

To reduce carbon dioxide emissions, carbon-free ammonia is being investigated as an alternative fuel for internal combustion engines. When injected into a low-pressure environment, liquid ammonia undergoes rapid vaporization (flash boiling) due to its low boiling point. In this study, the effects of ambient pressure and fuel temperature on spray morphology and tip dynamics of ammonia from a multi-hole injector were investigated under an injection pressure representative of port fuel injection. Visualization was performed in a constant volume chamber. Mie scattering was used to visualize the liquid phase distribution, while schlieren imaging was used to visualize the gas-liquid distribution. The results showed that, as the ambient pressure decreased and the superheat degree increased, the spray plumes from individual holes collapsed and merged into a single spray structure. The spatial distributions of liquid and gas phases remained similar, and the spray remained in the liquid phase up to the tip, regardless of ambient pressure and fuel temperature conditions. With increasing superheat degree, the spray region expanded near the nozzle and narrowed downstream. In addition, the penetration length increased. This increase was quantitatively correlated with the volume expansion caused by flash boiling.

Single-link used in rolling stock are an important component that connects the bogie to the car body. They trail and brake the car body according to the acceleration and deceleration of the bogie. The single-link buffer rubber used at both ends of the single-link is subjected to the high impact loads during bogie acceleration and deceleration, For this reason, the single-link buffer rubber is suitable for vibration isolation in railway vehicles, as it can withstand the high loads frequently applied to the vibration isolating rubber for rolling stock. The single-link buffer rubbers need to be replaced more frequently than normal vibration isolating rubbers and are required to have a longer service life. On the other hand, in order to extend the service life of single-link buffer rubber, it is necessary to understand the deterioration state of single-link buffer rubber used in actual vehicles and to develop a reproduction method that can reproduce the state of new products. However, such a method has not yet been established. Therefore, at first, we investigated the deterioration of single-link buffer rubber used in actual vehicles. As a result, it found that, while the overall tendency of single-link buffer rubber is to harden, it tends to soften in certain directions. Based on the results of this investigation, we conducted a combined test of heat aging test to harden the rubber material and fatigue test to soften the rubber material on a new single-link rubber buffer. The results showed that it was possible to reproduce the softening tendency in certain directions observed in used products.

Social robots are equipped with various sensors such as cameras, microphones, and tactile sensors, and realized rich interactions with children. Previous research has been reported that children felt like a close friend and were more likely to self-disclose than other humans when communicating with a social robot. Therefore, it is expected that it will be possible to build close relationships and develop trustworthy robots by introducing an active listening attitude to social robots, as if they were communicating with human. We have already developed Pupiloid, a speech-driven embodied listening system that expresses an active listening attitude using only the eyes based on the talker's speech. Pupiloid can give the talker a sense of attentive listening by generating nodding movements through eye movement and expressing interest through enlarged pupils. Furthermore, the effectiveness of Pupiloid’s listening attitude was demonstrated by a communication experiment with twenties participants. However, the impression that Pupiloid forms on children has not been investigated, and it is unclear whether the Pupiloid is effective on children. In this study, impressions of Pupiloid were investigated among a total of 110 children who participated in a hands-on experiment at an elementary school and a robotics event. The impression evaluation results demonstrated that children also formed favorable impressions of Pupiloid, with a high evaluation of its familiarity and sense of attentive listening.

We developed an automatic tagging system for design documents based on BOM (Bill of Materials) information. The purpose of the development is to facilitate efficient information retrieval during the design process and to support sharing expert knowledge in railway vehicle design. The system integrates two key natural language processing techniques. First, a BERT + CRF-based Named Entity Recognition (NER) model is utilized to accurately extract specialized railway component names from design documents. Second, an LDA(Latent Dirichlet Allocation)-based model is developed to infer potential tags by learning the latent structural relationships among component names from BOM data. A novel corpus construction method is proposed where BOM information, including hierarchical assembly structures, is converted into a Bag of Words (BoW) vector, enabling LDA to capture co-occurrence patterns and produce meaningful topic distributions. We conducted verification using 69 railway design documents manually tagged with reference component names. The evaluation, based on precision, recall, and F1 score, confirmed that the proposed LDA-based tag estimation model significantly outperforms the traditional TF-IDF approach with an F1 score approximately three times higher than that obtained by TF-IDF. These results demonstrate the effectiveness of leveraging BOM information to model latent structural relationships among component names. In conclusion, this technology can enhance the accessibility of relevant design information, contributing to a more efficient DX(Digital Transformation)-driven production environment.

In recent years, the proportion of elderly individuals in Japan has been increasing, accompanied by a rise in the number of patients with osteoarthritis. In particular, artificial joints, which are effective in treating knee osteoarthritis, commonly use Co-Cr-Mo alloys (CCM alloys) for their sliding surfaces due to their ability to suppress uneven wear. However, conventional CCM alloys pose a problem during sliding use, as they generate biotoxic metallic wear debris. To address this, high-carbon Co-Cr-Mo alloys (HC-CCM alloys) with improved wear resistance have been developed. Nevertheless, there have been few studies on the machining of HC-CCM alloy with end mills. This study investigates the high-speed machining of HC-CCM alloy using radius end mills, focusing on the influence of tool material and cutting speed on cutting performance. The results showed that both carbide and CBN tools reached their tool life limit within a short cutting length. In contrast, the binderless nano-polycrystalline cubic boron nitride (RcBN) tool exhibited excellent wear resistance, achieving a machining length of up to 7.0 km. These findings indicate that the RcBN tool is highly effective for machining HC-CCM alloy. Furthermore, excessive tool damage was observed at cutting speeds exceeding 15 m/s, suggesting that an optimal cutting speed is approximately 10 m/s.

Classical ballet and other forms of dance have both aspects of sports competition and artistic performance. Dancers require not only the technical skill to control body movements but also the expressive ability to effectively and impressively convey emotions and narratives to an audience. While physical movements can be measured and quantitatively evaluated, it is challenging to quantitatively assess people's subjective impressions of dance. Addressing this challenge, research aimed to quantify and objectively evaluate ambiguous Kansei information, handle nonlinear relationships, and specifically analyze skeletal movements in dance. To investigate the impressions ballet, a questionnaire survey was conducted. This involved using 18 pairs of Kansei words to evaluate nine sample videos, which allowed for the definition of decision classes. Subsequently, rough set theory was applied to analyze the relationship between ballet body movements and these impressions. Analysis focused solely on skeletal movements, deliberately excluding elements like background, music, and costumes that significantly influence overall impressions. OpenPose was utilized to detect human joint points and generate a BODY_25 Model. Dance characteristics were then defined using 8 attributes, each with two possible values to thoroughly analyze the correlation between body movement features and the impressions perceived by viewers. By utilizing rough set theory, the ability to extract patterns and regularities from data containing subjective human judgments is demonstrated. This enables an objective evaluation of the relationship between body movements and Kansei information such as "beautiful" or "elegant" derived from ballet's physical expressions. This approach proves that rough set theory's robust analytical capabilities can effectively analyze the nonlinear relationship between body movements and viewer impressions.

Finite elements analysis (FEA) has been widely used for evaluating the performance of products in practical engineering design processes. In the FEA of industrial equipment, such as large-scale motors and transformers, the accurate identification of material constants for iron core components is essential. This identification is formulated as an inverse problem based on mechanical properties. Surrogate multi-objective optimization has been widely employed to solve this problem, demonstrating high accuracy. However, the identified material constants are not always uniquely determined and may result in physically unrealistic values. From a practical perspective, it is crucial to obtain physically realistic material constants. To address this issue, this paper presents two key ideas: (1) the introduction of an in-plane isotropy constraint that reflects the structural characteristics of the iron core, and (2) the consideration of the sensitivity of objective functions with respect to material constants. The former effectively confines the search space to a physically plausible region, thereby enhancing both the physical realism of the identified constants and the accuracy of the FEA results. The latter contributes to improving the uniqueness of the identified values. These ideas are validated through a case study of an iron core component. The results of this study show the effectiveness in improving identification accuracy and ensuring the physical plausibility of the estimated material constants.

To reduce energy consumption in production systems, it is effective to evaluate energy-saving measures through simulations in advance. However, conventional methods require significant time and effort to construct and analyze multiple simulation models for different energy-saving measures configurations. Automating this process improves efficiency, but existing methods struggle with model complexity and reliance on human experience, risking the omission of effective measures. To address this challenge, we propose an automated approach using 1D-CAE models to systematically explore energy-saving configurations. Our method integrates engineering constraints in Alloy with Modelica-based simulation models to automatically generate and evaluate feasible system configurations. A key feature is a hierarchical constraint structure, where Modelica-dependent constraints are extended to define system-specific constraints, ensuring both syntactic validity in Modelica and physical relevance. We demonstrate this method using a 1D-CAE model of an injection molding machine, demonstrating a systematic exploration of energy-saving technologies.

Adaptive cruise control (ACC) is one of the critical elements of vehicle performance in the market. To ensure the quality of ACC performance, comprehensive evaluations that control both complex test scenarios that reproduce market driving conditions and vehicle behavior is required. However, it is difficult to evaluate all combinations of test scenarios using real test vehicles within limited development resources. Furthermore, it is necessary to determine complex Electrical Control Unit (ECU) parameters while considering multiple performance trade-offs. This paper proposes a new automatic screening and exploration system for ACC, incorporating Bayesian active learning (BAL). The proposed system automatically explores the worst conditions of ACC and the design space of ECU parameters for the improvement of vehicle ACC performance. This system consists of two automated elements: an automatic evaluation system and an automatic exploration system. In the automatic evaluation system, the behavior of ACC is automatically evaluated in real-time simulation using Real Car Simulation Bench (RC-S). Additionally, ACC sensor simulation is used to simulate various driving scenarios that may occur in the market. In the automatic exploration system, the worst condition screening evaluation of ACC performance and the exploration of the feasible region of design space for ECU parameters using BAL are conducted. As a result, it becomes possible to make the evaluation process more efficient through closed-loop evaluation, thereby improving ACC performance. In BAL, a Gaussian process model of ACC performance evaluated by RC-S is trained. Based on the posterior distribution of the trained Gaussian process model, the acquisition function is evaluated and maximized to generate new sampling points. In this study, an example of data comparison between RC-S and a real vehicle driving on a test course is demonstrated to show the effectiveness of the proposed system.

Shape design optimization of 3-dimensional solid structures must satisfy two key objectives: high mechanical performance and light weight. For machines in motion, time-response problems in design optimization are crucial for machine life and reliability, requiring optimal time-response characteristics depending on the given objective function. Time-response problems (e.g. the dynamic compliance minimization problem) arise not only in solid structures made of single materials but also in composite structures composed of heterogeneous materials. Heterogeneous composite structures can exhibit excellent mechanical behavior that is unattainable with single-material structures. For example, by using dissimilar materials with different thermal expansion coefficients, it is possible to control thermal displacement through interface shape design optimization between the two different materials. Shape design optimization is also beneficial for reducing the dynamic compliance of solid structures within a limited volume. Bimetals are a type of heterogeneous composite structure consisting of two different adhered metals. In recent years, metal additive manufacturing technology has rapidly advanced, allowing for precise fabrication of metal parts. This study aims to develop a gradient-based interface shape design optimization method for minimizing the dynamic compliance of 3D bimetal composites. First, we formulate the design problem, where the time-dependent dynamic compliance is set as the objective function to be minimized, with the time-response governing equation and the volume constraint serving as the constraint conditions. Then, we theoretically derive the shape gradient function (i.e., the sensitivity function) and perform the velocity analysis to obtain the optimal interface shape between the adhered metals. Furthermore, the effectiveness and feasibility of the proposed interface shape design optimization method are validated through design examples.

Plastic injection molding (PIM) is a major manufacturing technology to produce plastic products for high product quality and high productivity. The process parameters such as the melt temperature, the packing pressure and the cooling time affect the product quality and productivity, and thus it is important to determine the optimal process parameters. Recently, the pressure vibration injection molding (PVIM), which vibrates the pressure during the filling and packing phases, has attracted attention. The mechanical properties, the viscosity and the shear stress are mainly discussed through the experiment in the literature, but the period and amplitude of pressure vibration are rarely discussed. In this paper, the process parameters optimization in PVIM is performed for a thin-plate product. The warpage and the cycle time are simultaneously minimized for high product quality and high productivity. The numerical simulation in PIM is so intensive that sequential approximate optimization using radial basis function network is adopted. It is clarified though that numerical result that the PVIM makes the distribution of shear stress and pressure uniform, and thus the warpage is well reduced. Thus, the PVIM is an effective approach for warpage reduction and short cycle time.

This paper proposes the fluorescence response-based optical probing (FROP) method for the 3-dimensional measurement of precise products. Several 3-dimensional measurement methods exist, such as micro-coordinate measuring machines, confocal microscopy, and point autofocus microscopy. However, measuring precise products with small, smooth, and steep (3S) structures—such as die molds and optical lenses—remains challenging. In this study, we propose a new surface detection scheme that utilizes autofluorescence from the sample surface. Unlike reflected light, fluorescence is emitted over a wide angle. Therefore, the optical response from the surfaces of 3S structures can be obtained by exciting fluorescence at the measured surfaces. This paper first explained the principle of FROP. Next, the fundamental FROP signal was examined on surfaces tilted at different angles. The FROP successfully detected vertical and even overhanging surfaces, demonstrating its strong potential for 3-dimensional measurement. The principle of surface position determination was then verified through comparisons with conventional confocal microscopy for 2.5D measurements, and thickness measurement results were compared with micrometer results. These results revealed that the peak position of the differentiation signal in FROP coincided with the sample surface. Finally, a 3-dimensional 3S structure was measured. The results confirmed that vertical surfaces could be successfully measured using the FROP method, whereas conventional confocal microscopy could not measure them. Consequently, the performance of FROP for 3-dimensional measurement of precise products was validated.