Module-type deformable robots (MDRs), which combine multiple flexible modules to accomplish different tasks, offer superior adaptability over conventional fixed-morphology robots, making them ideal for rescue and exploration. However, existing control methods fail to fully exploit their potential due to insufficient consideration of external forces and mechanical characteristics of assembled structures. This study addresses these limitations utilizing a powerful structural analysis tool - a finite element method (FEM). As a fundamental step, we especially focused on demonstrating advantages of applying FEM to control of an MDR. We developed an FEM-based inverse analysis scheme using truss elements, along with a modular, deformable truss robot and its feedforward control system. The scheme takes target deformation and external forces as an input and solves a stiffness equation, estimating required actuation forces for each module. Embedded electromagnets and springs provide both passive and active axial movements with modules of the truss robot; the control system first performs inverse analysis and then converts actuation forces into current commands, considering internal and external force effects. Subsequent experiments showed the capability of the control system against simple two- and more complex four-module configurations even under significant self-weight influences. These results highlight the ability of FEM to control MDRs in various configurations and conditions, even without feedback loop, unlocking their full adaptability and functionality.

Estimation of stress–frequency distributions under in-service conditions is essential for assessing the structural integrity and fatigue safety of railway vehicle bogie frames, particularly at the design stage where actual measurement data are unavailable. This study aims to provide a simplified but effective approach to contribute fatigue life assessment or crack growth evaluation by generalizing and statistically characterizing stress–frequency distributions. A total of 113 measurement datasets were analyzed, obtained from eight types of bogies operated on nine conventional railway lines. Statistical evaluation showed a strong correlation between normalized stress, defined as the ratio to the maximum observed stress, and the logarithm of relative frequency normalized by total frequency. The slope of this relationship was found to follow a normal distribution, with a mean of 4.52 and a standard deviation of 0.66. Based on this result, two representative stress–frequency distributions with total frequencies of 10⁴ and 10⁵ cycles were assumed. For these distributions, equivalent running distances with respect to fatigue crack propagation were then evaluated in terms of percentile values. The findings indicate that the statistical representation of stress–frequency distributions provides a practical framework for estimating fatigue behavior in the absence of direct measurement data. Finally, potential applications of the proposed distributions were discussed, including approaches for estimating maximum stresses and effective utilization of the obtained statistical values in fatigue safety issues.

Noise emissions from construction and agricultural vehicles are a critical concern. As electrification progresses globally, noise from essential hydraulic systems, particularly pressure ripple in hydraulic-mechanical transmissions (HMT), becomes relatively dominant. Optimization for quiet designs of hydraulic systems is becoming more challenging due to the increased degree of freedom associated with the spread of model-based design (MBD), which often leads to NP-hard problems. Conventional methods - including those utilizing model approximation, gradient information, or probabilistic search - often fail to secure global optimality or require prohibitive computational costs. This study therefore proposes quantum annealing (QA), leveraging recent progress in quantum computing and the quantum mechanical tunneling effect for efficient global exploration, making it ideal for this class of NP-hard problems. This study verifies the feasibility of QA for reducing HMT pressure ripple. Using MBD data, the objective (pressure ripple) and constraint (efficiency) have been formulated. Continuous variables have been binary-encoded, and then the cost function suitable for the QUBO formulation required by QA has been defined via the penalty method. The QA-derived optimal solution has successfully located the Pareto front for the pressure ripple and the transmission efficiency, confirming the effectiveness of the method in exploring the optimal solution space. Future work will focus on enhancing mathematical model reliability and improving search efficiency through strategic SQA parameter adjustment based on engineering insights.

It is important to control the vibration of the car body and maintain good ride-comfort especially on high-speed railway. This paper focuses on the vertical vibration of car body. The vibration modes and intensities of actual running rolling stock were analyzed using the Tokaido-Shinkansen as an example. The results showed that the anti-roll damper between cars causes bending vibration peaks at a frequency lower than the first natural bending eigenvalue of the car body. Therefore, we developed a multi-body vehicle dynamics simulation model that can simulate a train set and clarified the vibration mechanism caused by the anti-roll damper between cars. Specifically, it was found that bendin g vibrations caused by the anti-roll damper between cars occur in the following two cases. The first case occurs when the following two conditions overlap. The first condition is that the even multiple of the half-wavelength of the longitudinal irregularit ies is close to the distance between the two bogies, and the front and rear bogies are vibrating up and down in the same phase. The second condition is that the odd multiple of the half-wavelength of the longitudinal irregularities is close to the distance between the vehicle centers, and the neighboring vehicles are vibrating up and down in the opposite phase. And the second case occurs when the following two conditions overlap. The first condition is that the odd multiple of the half-wavelength of the longitudinal irregularities is close to the distance between the two bogies, and the front and rear bogies are vibrating up and down in the opposite phase. The second condition is that the even multiple of the half-wavelength of the longitudinal irregularities is close to the distance between the vehicle centers, and the neighboring vehicles are vibrating up and down in the same phase.

A numerical simulation of conjugate heat transfer was conducted to analyze the thermal mixing phenomenon at a T-junction, with a specific focus on the reverse flow from the main pipe into the branch pipe. To investigate the causes of temperature fluctuations on the inner surface of the branch pipe, dynamic mode decomposition was applied to the simulated temperature field. The analysis revealed a dominant 66 Hz temperature fluctuation, with a corresponding Strouhal number of approximately 1, at the point where the main stream and branch pipe flow collide. A lower frequency fluctuation of approximately 30 Hz was observed at the right corner. In the reverse flow region along the right wall of the branch pipe, fluctuations of 5.5 Hz and 0.29 Hz were also detected. These findings suggest that the vortex roll-up generated in the central part of the branch pipe collides with the right corner, leading to relatively low-frequency temperature fluctuations on the inner surface. Based on this analysis of temperature fluctuation factors, the possibility of reducing the risk of thermal fatigue damage in the branch pipe was investigated by altering the merging angle of the T-junction. When the merging angle was changed to 45 degrees, the root mean square values of both the inner surface temperature and the fluid temperature near the branch pipe wall decreased compared to the 90-degree case. This change also led to a reduction in the maximum range of temperature fluctuation on the inner surface of the branch pipe, a crucial indicator for fatigue assessment. The results suggest that the amplitude of stress at the upper end of the thermal sleeve installed in the branch pipe can be mitigated by changing the merging angle of the T-junction.

This study applies a hybrid finite element-material point method (FEM-MPM) framework to the simulation of pneumatically actuated soft robots undergoing large deformation and contact. Internal forces and pressure boundary conditions are evaluated using an explicit total Lagrangian FEM, while contact—including self-contact—is handled robustly via an MPM formulation with P2G/G2P mappings on a background Eulerian grid. Interactions with external bodies employ a frictional contact model, and self-contact is treated succinctly through MPM’s inherent non-penetration. The cavity volume enclosed by the FEM boundary mesh is updated incrementally via the divergence theorem, and internal pressure is prescribed through an isothermal equation of state, enabling an approximate coupling between deformation and internal pressure while avoiding the need for sophisticated full-scale fluid–structure interaction analysis. As verification, we analyze a self-contacting bending actuator and a four-finger soft gripper assembled from the actuator, reproducing grasping sequences with complex contact in a consistent manner. We further show that global deformation states can be identified from pressure time series alone, providing a basis for numerical analysis and data acquisition toward self-sensing soft robots using only injected volume and actuation profiles. Finally, a case study that systematically varies the initial arrangement and injected air quantity reveals―summarized as a phase diagram―that grasp success depends strongly on spatial configuration and actuation conditions. This application indicates that the hybrid FEM-MPM workflow can serve as a practical foundation for design exploration and future closed-loop control.

Model predictive control is applied to real plants because the control algorithms are easy to understand when designing control systems and can be applied to multivariable systems. It has been applied particularly to petroleum and chemical plants, and recently, its application has been expanded to mechatronics systems such as robots. As wheel-type inverted pendulum robots are compact and have high movement performance, there is a demand to follow a target trajectory while stably standing upright. So, there have been reports of model predictive control applied to wheel-type inverted pendulums. However, the effectiveness of these methods has been verified by simulation, and verification using the actual robot has not been performed. In addition, nonlinear model predictive control is mainly applied to deal with constraint conditions for the purpose of avoiding obstacles on the trajectory. One issue with nonlinear model predictive control is that in online control, the solution for nonlinear optimization cannot be calculated within the sampling time. In fact, it is often sufficient to calculate the target trajectory to avoid obstacle offline. For such objects, linear model predictive control has the advantage that the structure of the control is clearer and it is easier to adjust the control on-site, rather than nonlinear one. In this paper, we created a wheel-type inverted pendulum robot, and describe the results of applying trajectory tracking control based on model predictive control. Here, linear model predictive control is applied to achieve target trajectory tracking and state feedback control is applied to stabilize the wheel-type inverted pendulum robot against disturbance, and these control inputs are added together to achieve control. In addition, we describe the issues and solutions for parameter identification, target trajectory setting, and control design, which are important in the practical application.

In order to clarify the effect of surface hardening layers induced during machining processes such as cutting and grinding on the wear resistance of austenitic stainless steel SUS316L, a surface hardening layer was intentionally induced on SUS316L through lathe cutting. A ball-on-disk friction wear test was then conducted on the hardened SUS316L using an alumina ball with a load of 9.8 N, total sliding distance of approximately 7.9 m to 31 m. The results showed that the samples with a surface hardening layer had a significantly improved volumetric wear rate (up to 37% improvement) compared to those without a hardening layer. To understand the reasons for this improvement, the observation of cross-sectional microstructure, hardness tests, and profile scanning of the test specimens after the wear test were conducted. These observations indicated that plastic deformation suppressed by a harder surface layer on hardened sample, and inducing stable oxidation layer on the worn track which resulted in a reduction in the amount of wear. This suggests that intentionally inducing and retaining a surface hardening layer during manufacturing, especially during machining, can potentially enhance the wear resistance of SUS316L products.

Laser-based powder bed fusion of metals (PBF-LB/M) is one of the additive manufacturing (AM) techniques that enable the building of complex structures with high-precision. In particular, injection molds and die casting dies are products to which the advantages of PBF-LB/M are applied, such as improving the cooling performance by installing flexible cooling channels and suppressing the insufficient resin filling and the gas burning by incorporating the permeability through the porous structures. In this study, a new laser scan strategy is proposed to further improve the functionality of injection molds by building the microtubes concentrically. The timing of laser irradiation and the operation of galvanometer scanner were investigated, and the influence of microtube size on the actual laser scan speed and heat input was evaluated. In addition, the effect of correcting the amount of heat input on the accuracy of microtubes was also investigated. As results, the actual laser scan speed decreased as the number of control points on the scan path increased, and the decrease of actual laser scan speed caused the blockage of microtubes due to the increase of heat input. Furthermore, the heat input increased as the microtubes became smaller due to the acceleration and deceleration of the laser scan. Correcting the laser power according to the actual laser scan speed made it possible to form smaller microtubes. With a spot diameter of 0.2 mm and heat input corrected to 0.457 J/mm, the square microtube exhibited reduced blockage below 0.3 mm and formed 0.1 mm.