We publish "Toyota Technical Review (TTR)" with the aim of fostering partnerships through technology introductions to stakeholders and recording Toyota's technical assets.
Latest NumberVol.71
Date of Issue
- (Japanese Version) November 25, 2025
- (English Version) March 23, 2026
Contents
Special Feature: Application of Vehicle Data-Driven Diagnostics Technologies to Address Mobility-Related Issues
Part 1 Overview of Diagnostics Technologies Utilizing Vehicle Data
Part 2 Driving Diagnostics Technologies for Zero Accidents and Carbon Neutrality
- Driving Safety Support Using KINTO Connected Driving Trainer
The connected driving trainer (CDT) feature was developed with the aim of enhancing drivers' safe and environmentally friendly driving skills as part of Toyota's initiatives to help reduce traffic accidents and achieve carbon neutrality. It consists of two functions: a driving coach function that visualizes driving habits and an upgrade recommendation function that proposes optimum upgrades to the driver. These functions use onboard sensors to provide the type of diagnostics services that only an automaker is capable of offering, such as highly accurate and rapid diagnostics of the driver's turn signal operation and lane sense. - Driving Safety Support Using Rental Vehicle Driving Monitoring
Starting in 2022, Toyota has been participating in the Okinawa Yuimaru Project, a data-driven initiative to help reduce the number of accidents in Okinawa Prefecture. As part of this project, a driving monitoring app was developed for the in-vehicle multimedia units of rental vehicles to help raise the safety awareness of rental vehicle users. This article introduces the main functions of this app, namely the safe driving status notification function and the dangerous location warning function, the results of a demonstration test carried out with Toyota Rent a Car Okinawa starting in June 2024, detailed observations of these results, prospects for further development, and issues for enhancing app functionality. - Estimation of Cognitive Function Decline
This article describes the development of a system capable of estimating cognitive function decline in drivers from vehicle data obtained while driving. First, Toyota partnered with research institutes to construct a labelled mild cognitive impairment (MCI) driving data set that includes both cognitive function assessment results and vehicle data. Next, a four-step cognitive function decline estimation method was proposed based on vehicle data obtained from driving scenarios: driving scenario extraction, feature quantity extraction,estimation of cognitive function decline in each driving scenario, and the calculation of representative values for the estimated values in multiple scenarios. Finally, the accuracy of the proposed method was verified by cross-validation. It was confirmed that this method is capable of estimating the cognitive function decline of drivers to a statistically significant degree. By increasing the chance of early MCI discovery, this system has the potential to help maintain the quality of life (QOL) of elderly drivers and contribute to traffic safety. - Support for Elderly Driver Training Courses
This article describes the development of a system that allows elderly drivers to gain self-awareness of their own driving capabilities and confirm the results of diagnostics for those capabilities. This system is implemented as a part of actual vehicle driving training for elderly drivers carried out at driving schools for the purpose of extending people's active driving age. This development was carried in partnership with driving schools, which helped to select the diagnostics items and carried out demonstration tests of the system. The system analyzes driving operations and vehicle behavior using vehicle data, and carries out driving diagnostics based on vehicle trajectories obtained using high-precision vehicle position measurement and training course maps. After the training course, the participant is given a driving diagnostics report and the results are also made available on a website for the participant's family. In this way, the participant and participant's family can understand for themselves the mental and physical condition of the participant, which can then be used as the basis for decisions about the necessity for further driving training or the like.
Part 3 Vehicle Diagnostics Technologies for Even Greater Peace of Mind and More Efficient Vehicle Inspections
- Prediction of Deterioration of Vehicle Components Requiring Maintenance
Vehicle components and consumable items that require maintenance are replaced based on guidelines such as distance driven and elapsed time. This article uses the example of engine oil to describe the outline of a system that utilizes vehicle operation and driving data to predict the state of deterioration in accordance with usage conditions. This approach can be used to optimize part replacement, thereby helping to promote carbon neutrality and reduce dealer workload. The developed system has already been adopted as part of the Connected Car Care feature on vehicles available through the KINTO subscription service in Japan, and Toyota has plans to both roll out the system globally and expand the number of applicable components and consumable items. - Support for Rental Vehicle Inspections by Minor Collision Detection
Toyota has developed a minor collision detection system using driving data from connected cars to help alleviate the workload of businesses when vehicles are damaged in a collision. In the development of this system, a method was proposed that estimates acceleration caused by driving operations using a long shortterm memory (LSTM) model and then detects minor collisions by subtracting that acceleration from actually measured acceleration values. This method was verified from data obtained over half a year from around 1,000 taxis operating on the outskirts of Tokyo. The results found that this system is capable of detecting approximately 90% of accidents that cost 200,000 yen (around 1,300 USD) or more to repair, while also estimating the location of the damage from the acceleration direction. Another demonstration test using 850 rental vehicles in Okinawa showed that the number of false positives caused by local characteristics could be reduced by post-processing. Work is currently under way to enhance this system to enable even more detailed collision detection while minimizing false positives. - Detection of Signs of Wheel Detachment
With accidents involving tire and wheel detachment attracting increasing attention, Toyota has developed and verified the effectiveness of an algorithm that detects signs of wheel detachment using wheel speed sensor information from connected cars. In partnership with dealers, data from approximately 550,000 vehicles over around five months was analyzed to verify the situation on actual vehicles. The effectiveness of the algorithm in detecting signs of detachment such as lug nut looseness was confirmed. However, since the developed algorithm also detected a wide range of abnormal conditions as well as lug nut looseness, further work is needed to enhance its accuracy. In the future, Toyota is aiming to help realize an even safer vehicle-based society by enhancing and implementing this detection algorithm in the real world.
Part 4 Infrastructure Diagnostics Technologies for More Efficient Road Management
- Support for Paved Road Surface Inspections
This research developed a system that estimates the state of flatness and ruts on paved road surfaces using vehicle probe data obtained from connected cars. First, flatness estimation was carried out by analyzing input data based on fluctuations in vehicle wheel speed and generating an indicator that quantifies the state of paved road surfaces. After confirming the validity of the indicator values by comparison with road repair data, verification tests were carried out in partnership with actual road management authorities. These tests determined that the developed indicator achieved a correlation coefficient of R = 0.65 compared to the maintenance control index (MCI) commonly used by road administrators. Next, the state of ruts was estimated by evaluating changes in the road surface reaction force using vehicle lateral acceleration. The resulting indicator is highly accurate and recorded a correlation coefficient of R = 0.75. These indicators have the potential to enhance the efficiency and reliability of road inspections and advance the feasibility of real-world adoption of this system. - Support for Traffic Safety Measures
This article discusses the utilization of vehicle probe data obtained from Toyota's connected cars in road safety measures that are designed to help reduce traffic accidents. Vehicle probe data can be statistically processed to create quantitative frequency indicators for events such as sudden braking, stopping at intersections without a traffic signal (unsignalized intersections) , speeding, and so on. These indicators can then be used to identify hazardous locations that are difficult to detect by conventional subjective surveys, as well as to help propose and determine the priority order of safety measures. In addition, data from before and after the adoption of a safety measure can be compared to enable objective verification of its effectiveness. This technology has also been utilized in field demonstrations and other aspects of real-world traffic safety measure development, helping to enhance the efficiency of these measures. - Queue Length Estimation
As the use of vehicle probe data from connected cars expands, Toyota is working on ways of using this data to address social issues. As part of this approach, this article describes Toyota's measures for mitigating traffic congestion, which is a major vehicle-related issue. Since many roads administered by local government bodies are not equipped with roadside sensors, measures for mitigating traffic congestion generally utilize local visual-based surveys. However, because such local surveys can only be carried out for limited periods due to cost and time constraints, surveys might not be capable of identifying the state of traffic congestion accurately. In response to this concern, Toyota has developed a data acquisition system capable of replacing local surveys by using vehicle probe data to simulate queue length, which is commonly used as an indicator for traffic congestion mitigation by local government bodies. This article outlines this technology and describes an example of its use for traffic congestion mitigation in Toyota city, Aichi Prefecture.