Development of Other Vehicle Behavior Model for BEV Thermal and Dynamic Evaluation Simulator

Masahiro Nojima･Masaki Morita･Takuji Horimoto (Toyota Motor Co.)

In BEV development, many new thermal induced power development issues have emerged. So, evaluation is necessary, but currently there were problems with man-hours and other constraints for on-site evaluation, making it desirable to utilize a driving simulator. However, it was not possible as traffic flow had not been established.
In this study, we developed an algorithm to extract big data from other vehicles, taking into account the traffic dynamics and interactions with our own vehicle. This algorithm accurately reproduces real-world conditions, enabling implementation and evaluation in a driving simulator.

Study on development efficiency and system cost reduction through BEV charging service platform

Midori Sugiyama･Yoshihiro Sakayanagi･Masato Ehara･Takahiro Hirano･Tsubasa Otohata (Toyota Motor)

To achieve carbon neutrality by 2050, we are developing a management system that integrates energy management with mobility services. The purpose of this work is to develop a platform that enhances development efficiency, shortens lead times, and reduces operational costs through using of APIs, communication protocols, and serverless architecture. Additionally, we summarize effects and insights gained from demonstration experiments.

Development of Logistic Demand Estimation Process and Supplementary Methods for Predicting Future Changes in Truck Demand.

Kenta Shintoku (Kozo Keikaku Engineering)･Nobumasa Ohashi･Junichiro Nitta (ISUZU ADVANCED ENGINEERING CENTER, LTD)･Ryoko Maeda (Kozo Keikaku Engineering)

In this study, we develop a future demand forecasting process for trucks based on public statistical data such as the Logistic Census. To address data limitations, we incorporate supplementary estimation methods based on prior research, allowing evaluation demand by regions, transportation modes and categories. Furthermore, through scenario analyses that takes into account the expected progress of modal shift, we provide a framework to capturing future structural changes in logistic demand. This study demonstrates the effectiveness of combining structured forecasts with data-based supplementation to support strategic planning regarding future trends in logistic demand.

Design and Evaluation of a Virtual Traffic Light (VTL) Control Algorithm Adaptive to Traffic Demand

Keita Sakai (Toyo University Graduate School)･Hironori Suzuki (Toyo University)

By using vehicle-to-vehicle (V2V) communication technology, Virtual Traffic Lights (VTL) can control intersection traffic flows more efficiently than conventional physical traffic lights. In this study, we design a novel VTL control algorithm that can flexibly respond to fluctuations in traffic demand. The proposed algorithm is implemented in a traffic flow simulator, and its effectiveness in improving traffic performance is evaluated through multiple scenarios.

Trajectory Prediction of Traffic Participants in Interaction Scenes at Signalized Intersections

Quy Hung Nguyen Van･Heishiro Toyoda (Toyota Motor)･Cui Xiongyi･Rosman Guy (Toyota Research Institute)･Kimimasa Tamura (Woven by Toyota)

In this paper, we present our approach to predicting the trajectories of traffic participants at signalized intersections in urban areas in Japan. Based on measurement data from sensors (cameras, lidar) installed at intersections, we focus on interactions between traffic participants (particularly, crossing pedestrians, bicycles, and vehicles turning right and left), and investigate the feasibility of learning and predicting trajectories in situations where real interactions between traffic participants occur by utilizing advanced deep learning models.

A Robust License Plate Recognition System Against Environmental Changes Using Vision-Language Models

Kota Shinjo･Shintaro Yoshizawa･Yuto Mori (Toyota Motor)

License plate recognition systems have been used for parking lot management and traffic monitoring, but it is desired to expand the scope of their use to wider fields, such as factory premises and smart cities. To this end, we have developed a technology to realize more advanced license plate recognition with high accuracy by utilizing a visual language model that can flexibly respond to a variety of images.