Infrastructure maintenance and management in Japan, which is said to be an "advanced" nation, is a serious problem due to the lack of financial resources and experts. On the other hand, in recent years, sophisticated image recognition has become possible because of the development of technologies such as deep learning, and the increased refinement of camera functions of smartphones, which are widely used worldwide.

In this research, in cooperation with road managers of seven municipalities, we conducted real-time detection of road surface damage by deep learning, and attempted to extract criteria for repair correspondence decision for each local government by random forest method. As a result, it was possible to detect road surface damage with a detection rate of 87% using only general smartphones, and it was possible to catch a glimpse of the difference in standards for maintenance level correspondence between local governments. With this result, inexpensive and simple infrastructure inspection becomes possible, and there is a possibility of breakthrough in various areas suffering from financial resources and expert shortage.

The visibility of the pedestrian by conventional road lighting is known to depend on the relationship between the driver’s position and the pedestrian’s position. Pro-beam road lighting is a foundational solution for enhancing the visibility of pedestrians crossing the road. In the present study, we developed a proposed pro-beam luminaire that uses existing LED lamps and is designed to adjust the pro-beam road lighting distribution to achieve an ideal distribution. The visibility performance of the proposed pro-beam road lighting was evaluated by 16 young participants seated in a car on a test track. A field experiment found the visibility performance of the proposed pro-beam lighting in the longitudinal direction of the road to be higher than that of conventional road lighting. In addition, the present study revealed that vertical illuminance at the height of 0.8m strongly correlates with the average subjective visibility value. According to this relationship, two light distributions for the pro-beam road lighting were examined to improve the weaknesses of the proposed pro-beam road lighting.

This study reports the results of an experiment with a driving simulator where drivers were exposed to a simulated earthquake while driving on an urban expressway. Some drivers were given guidance before the earthquake on the recommended driving behavior during an earthquake in line with the official guidelines of the expressway company. We analyze the impact of this information and generally report on how drivers reduce speed as a function of traffic density and the seismic intensity during the simulated earthquake. We report on the overall level of speed reduction as well as the average and maximum deceleration. This information will help to build car-following models under emergency situations and to understand the accident likelihood during earthquakes. It will also help improving the guidance given to drivers. This appears to be of need since we find that drivers can have different interpretations regarding the meaning of a “stop” message. Whereas some drivers would aim to stop without much emergency on the hard shoulder, others would aim to stop rapidly, independent as to whether a safe stopping space could be found.

In this study, narrow roads are divided into 13 classifications by their differences in investigation method, size of traffic volume / traffic density, and route type. The results of the estimation model are compared, in order to show the possibility to estimate the narrow street traffic volume by several classifications. As a result, the following points were clarified. In 600 units/h or less, the mesh traffic-density model of a thin road is suitable, because of a high coefficient of determination. In 600 units/h or more, the traffic-volume model of “600 units/h or more” is suitable. Appropriate estimation accuracy was obtained by the combination method (traffic-volume and mesh traffic-density models of a thin road). The obtained results showed that traffic volume of narrow roads could be estimated by taking different methods in two divisions with 600 units/h as the limit.

The ratio of bicycles accidents at junctions in Japan is quite high. Both-way running of bicycles at junctions, which is unique system in the world, is focused on as a reason of this phenomenon. Bicycle lanes, which promote bicycles running on left side of the carriageway, have been introduced to solve this problem by realizing one-way bicycles running system. In this study, in order to confirm the effects on safety by bicycle lane system on residential streets in Kanazawa city, the approaching behavior of bicycles is focused on. Bicycles’ behavior and vehicles’ speed before and after the implementation of bicycle lanes at five junctions using video survey are analyzed, and bicycles’ behavior at three junctions with different accident records and condition of bicycle lanes is compared. As results, the ratio of wrong way approaching of bicycles is decreased by the implementation or existence of bicycle lane system which promotes keep left running of bicycles on the approach streets.

Urban transportation system consists of multiple travel mode operated by many agents. Thus, improvement of transportation hub are required to achieve comfortable urban transportation network. This research focused on comfortable lightening environment of bus stop. We implemented outdoor experiment to examine the effect of the degree of illuminance and the color temperature of lightening around bus stop on comfortableness of users. 15 experimental condition, which were consisted by 5 illuminance conditions and 3 color temperature conditions were prepared. Participants of experiment were requested to answer questionnaire at experiment site.

As a result of experiment, respondents showed high sense of comfortable, sense of safety, visibility of the space, sense of calm and satisfaction at high illuminance conditions. Besides, they showed high sense of safety and calm at low color temperature conditions. These effects were significant at low illumination (5lx-30lx) experiment conditions. Results of implied that illuminance of comfortable bus stop are required more than 30lx. In the case of low illuminance (under 30 lx), low temperature color of lightening could make people comfortable.