A new statistical model can be used to prevent accidents involving children
Insight
A research team led by Associate Professor Kojiro Matsuo of the Department of Architecture and Civil Engineering at Toyohashi University of Technology and Professor Kosuke Miyazaki of the Department of Civil Engineering at National Institute of Technology – Kagawa College created an effective tool. identification of road junctions with a potentially high risk of accidents involving children. Big road traffic data, past accident data, intersections of traffic structures and land use etc. in addition, a statistical model based on unique geo-informatics data for school commuting routes and school walking groups maintained by Toyohashi City can be used to prevent accidents. by identifying potential high-risk areas for such accidents where no accidents have occurred.
Details
In Japan, children often walk alone from about the age of seven (1St and 2no primary school classes). Although this is somewhat unusual from an international perspective, it is considered to contribute to the health and development of children. It is therefore vital to ensure that this can continue, and therefore increasing the safety of children’s travel is an essential requirement.
Proper identification of the areas and locations where road safety measures should be implemented is an integral part of road safety management to improve the safety of children on the move, but identification of these points is complicated by the rarity of such incidents. a large number of places where children pass. The potential risk at each location should be quantified, effectively identifying where additional measures are needed.
Thus, the research team developed a method using a statistical model to effectively identify potentially dangerous intersections in terms of traffic accidents involving children.
Associate Professor Matsuo, as Principal Investigator, explained: “This statistical model has two main features. First, it uses empirical Bayesian computations to incorporate the average effect of traffic environmental conditions such as traffic volume, intersection pattern, and land use status derived from large traffic data—appropriately balanced with past crash data.
The second is to incorporate information on the amount of children’s trips with traffic conditions, using unique geo-computerized data on Toyohashi city’s school walking routes and walking groups.
Finally, the fact that it identified seven or more potential danger points proved that it was more efficient and effective than methods based solely on past accident data.
In addition, we have confirmed the existing policy of “first implement road safety measures on school routes” because Japanese children tend to use familiar routes not only to and from school, but also for all daily walks, but accidents along these routes are increasing. »
Development context
Associate Professor Matsuo added: “This research would not have been possible without the geo-computerized data on school routes and walking groups provided by Toyohashi City.
I worked as a security consultant for school commuting routes in Toyohashi City, and as an administrative initiative, I planned the effective management and use of school route data in 2015. After proposing the use of computerized geo-data and data structure, it has been officially adopted since 2016. I had no idea at the time that this data would be useful for my research, but I am glad that the measures introduced six years ago are still being used. I hope that in the future other municipalities will also encourage the use of geocomputerized data with school routes. »
Future prospects
Through this research, we developed a method to identify potentially dangerous locations based on objective information obtained from various data sources. However, real traffic is complex and data is only one part of it. The research team wants to encourage research into how best to combine publicly held subjective data, such as close calls, with objective data.
