In the non-local eddy diffusivity model recently proposed, the energy spectrum was used, and it is not clear whether the model can be applied to inhomogeneous turbulence. In this study, the non-local model was improved using the scale-space energy density instead of the energy spectrum. The scale-space energy density and the two-point velocity correlation were examined using the direct numerical simulation (DNS) data of homogeneous isotropic turbulence, and a new non-local eddy diffusivity model was proposed and validated using the DNS data. In particular, the temporal behavior of the three-dimensional non-local eddy diffusivity was improved compared to the previous model.

The gradient diffusion approximation for the turbulent scalar flux is not valid and counter-gradient diffusion occurs in the turbulent premixed flame. Many processes due to the interaction between the turbulence and the chemical reaction are related to the counter gradient diffusion in the turbulent premixed flame. In this study, we focus on the effects of heat release to the turbulence. To discuss it, we devise and implement the direct numerical simulation of the turbulent Rayleigh flow (TRF), which is generated by isotropic turbulence flowing into Rayleigh flow. We show that the counter-gradient diffusion of the turbulent heat flux occurs in TRF and discuss its mechanism based on the transport equations of turbulent statistical quantities. It is shown that released heat amplifies density-internal energy correlation resulting in the prominent contribution of the pressure gradient effect, which brings counter-gradient diffusion.

In recent years, semi-closed circulating sea surface aquaculture in calm coastal waters has garnered attention. In this study, the sloshing of a closed rigid cage tank was analyzed using the particle method, concurrently examining changes in the internal free water surface shape and flow field due to sloshing. The results revealed significant nonlinearity in the free water surface shape near the 1st and 2nd modes of sloshing during the excitation period, leading to an overall increase in the internal water particle velocities. Overflow of water particles within the cage resulted in the resonance point of the upsurge height not appearing at the 1st mode but transitioning toward the longer period side. Additionally, a phase difference between the waves due to sloshing inside the cage tank and the displacement of the cage wall was observed. Our future work involves constructing a model that combines the particle method and finite element method to analyze the motion of closed flexible membrane fish cage in waves.

In Kuji Bay, a seawall at the bay entrance will be constructed to counter tsunami, while cage aquaculture has developed in the bay’s calm area, making it necessary to further investigate changes in the flow field and aquaculture waste dispersion. This work developed a coupled physical and geochemical model that includes an aquaculture submodel. Results revealed minimal impact on Kuji Bay’s current circulation and the flow around the fish farm due to the seawall. Concerning aquaculture waste dispersion, unconsumed feed concentrated more near the fish farm seafloor, and increased residence time in the bay led to longer settling times for feces.

Data-driven wind sensor correction based on wind observations and artificial neural networks (ANN) faces practical challenges due to the high temporal and equipment costs of collecting instantaneous wind speed data. This study aims to reduce the need for wind observations using a CFD database. In this report, we propose a two-stage prediction framework-based correction method and investigate the optimal construction of ANNs using the CFD database as training data. The results show that using measured data at one location and the CFD database as training data can improve the quantity and generalization performance of existing measured data, thereby enhancing the efficiency and accuracy of sensors at different locations.

One of the effects of turbulent helicity on the Ekman boundary layer is to enhance the winding of the Ekman spiral. Using the analytical method of Yokoi and Brandenburg (2017), we calculated the specific change in velocity and considered which term of the Reynolds stress is responsible for the effect. Finally, we concluded that the turbulent helicity term does not necessarily directly cause changes in the velocity field, but the viscous stress term does.

Turbulence is ubiquitous in astrophysics, geophysics, fusion plasmas, and other science and engineering fields. Turbulence in general is not in equilibrium state. Turbulence quantities, such as fluctuation energy and helicity, are not homogeneous in space nor stationary in time. In the present study, we consider the effects of non-equilibrium magnetohydrodynamic (MHD) turbulence on the generation mechanisms of global magnetic field (dynamos). In the present theoretical framework: the multiple-scale direct-interaction approximation, the analytical expressions for the turbulent transport coefficients are given in terms of the response functions and spectral functions of the turbulent statistical quantities. In this work, we focus our attention on to the role of the cross-interaction responses on the dynamo transport: the velocity fluctuation response to the magnetic perturbation and the magnetic-field fluctuation response to the velocity perturbation. It is shown that the non-equilibrium properties of turbulence may enhance the alpha-type dynamo in the co-existence of the turbulent cross helicity and turbulent kinetic helicity. Such an effect is confirmed by a series of direct numerical simulations (DNSs).

There are two increasing shared mobilities in Japan: car-sharing, which is expanding in urban areas, and demand-responsive transit, which is being introduced in rural areas. Rural MaaS proposed in this manuscript hypothesizes that combining these two shared mobilities can provide mobility services in rural areas without relying on private cars. This manuscript develops a framework for examining Rural MaaS based on previous cases and studies and discusses the possibility of the hypothesis.

In addition to technical and legal issues, improving the acceptance of automated mobility services, such as automated driving, has been pointed out as a challenge for their social implementation. In this study, we conducted a questionnaire survey of participants in a student contest on automated mobility services before and after the contest, asking them about their expectations and awareness of issues related to social implementation, and analyzed changes in their answers. As a result, it was found that there was a change in the desire for social implementation from limited space and environment and that there was a tendency for the participants to place a higher value on the need for government investment in R&D and infrastructure.

This research proposes a new method of vehicle localization using an environmental magnetic field (EMF). EMF consists of the Earth’s geomagnetic field and magnetic field induced by manmade objects. EMF has local fluctuations that can be paired with its coordinate position and is time-invariant within the practical scale, so it has the potential to be used as a localization method. In this paper, an EMF map was generated by Gaussian process regression using magnetic field data collected from our experiment vehicle. Then a simulation was done to verify the localization using the Monte Carlo localization algorithm. The proposed method was able to generate an EMF map resembling the original magnetic field measurements, and the algorithm successfully localized the vehicle. The results also showed that the magnetic markers may increase the localization accuracy.

The objective of this research is the social implementation of an automated valet parking system (AVPS) with a cooperative system. Challenges to social implementation include unclear roles and reform of parking infrastructure. In this paper, we conduct a virtual simulation experiment focusing on structural changes in parking lots, aiming to reduce waiting time and improve the efficiency of parking spaces.

In this paper, a novel perception framework is presented for 2D and 3D object detection, based on sensor fusion of cameras and Li-DAR. While camera images provide abundant environmental features, they lack depth information. Conversely, Li-DAR point clouds offer accurate depth information, which however, are sparse in nature. Recognizing the complementary nature of each sensor’s strengths and weaknesses, an unsupervised depth completion network to enrich information from both sensors is used. This enhanced data is then utilized for performing 2D and 3D object detection tasks using a state-of-the-art detection network. The proposed framework is validated on KITTI data set, and experimental results demonstrate notable improvements in both 2D and 3D tasks when compared to baseline results.

This paper introduces a novel two-stage approach for monocular 3D object detection and tracking. The methodology centers on an advanced depth estimation technique that produces high-quality depth maps, enriched with detailed spatial infor-mation. These depth maps, when combined with image data, serve as the input for a distinctive detection network that incor-porates multiple CenterNet heads. Further enhancing the network’s efficacy, deformable convolution networks are integrated into the backbone. The results demonstrate the method’s proficiency in monocular object detection. Leveraging the robust detection outcomes, the study extends its application to object tracking, showcasing its commendable performance in this domain.

Traffic safety at unsignalized intersections is ensured by priority rules among road users. However, despite legislation supporting pedestrian priority, vehicles do not always yield to pedestrians at such intersections. Particularly on urban streets where on-street parking and land-access are likely to be around intersections, interactions between vehicles and pedestrians are quite complex. In this paper, we conduct a trial analysis of vehicle-pedestrian interactions at unsignalized intersection on urban street using multiple object tracking, which has been rapidly developed in these years. Applying YOLOv8 to a video recorded from lamppost, pedestrian and vehicles trajectories are obtained and analyzed.

In this study, we use a micro traffic simulator to simulate the effect of changing the green time split for an intersection in Kikuyo-machi, Kikuchi-gun, Kumamoto Prefecture, and then the split change is implemented with the cooperation of the prefectural police. We compare the preliminary prediction results from the simulation and the observe values when changing the green time split. After that, we adjust the green time split using link travel time data obtained from the Google Maps API, and analyze the effect of changing the split by the link travel speed for five different scenarios.

Traffic crashes on expressways tend to be more serious when vehicle speeds are high, and prevention of such crashes is important for road safety. Many existing studies have investigated the factors on such crashes based on the sections where crashes frequently occurred. On the other hand, characteristics of the sections where crashes are prevented are not well understood. In this study, therefore, a comparative analysis of multiple sections with similar road alignment but different crash rates is conducted. By integrating and analyzing crash, road alignment, and detector data of the Tokyo Metropolitan Expressways, characteristics of the sections where crashes seemed to be prevented are investigated.

It is necessary to prevent pedestrians from entering after an alarm by their own decision to improve the safety of pedestrian behavior at level crossings. This study proposes a new interface that presents visual information to notice the timing of the crossing alarm, and therefore it encourages pedestrians to change their behavior. This study conducted a pedestrian experiment on a virtual environment, and the results suggested that the proposed interface will prevent pedestrians from entering the crossing after the warning by making their decision earlier to enter or stop at the level crossing.