top of page

Dangerous Driving Among the Elderly

Identifying the risk factors of cognitive decline in aging drivers

white car blurred speed motion


  • Japan's aging population is leading to an increase in dementia cases and traffic accidents involving elderly drivers, presenting a significant social challenge. The lab, using extensive telematics data, is collaborating with Mind Foundry to develop an algorithm that can detect signs of cognitive decline in drivers to help reduce accidents. This innovative solution, now in the validation phase, has potential for global application.



  • Japan is leading the world in terms of an aging population. As of 2020, 4.7% of the population, or 6 million people, are suffering from dementia. This number is expected to rise to 7.3%, or 8 million people, by 2040 due to the advancing aging of the population, as projected by the Cabinet Office. Moreover, traffic accidents caused by elderly drivers are becoming a significant social issue.

  • Aioi Nissay Dowa Insurance held 1.8 million telematics insurance policies as of September 2023, accumulating driving data equivalent to 90 billion miles (equivalent to 3.65 million Earth circumferences). This data includes driving behaviours and accident data of elderly drivers.

  • Traffic accidents are most frequent among the young and the elderly. The former group's primary causes are lack of skill and reckless driving, while the latter is mainly due to a decline in cognitive functions. Currently available telematics insurance primarily addresses the former group.


  • By utilizing vast amounts of telematics data and advanced machine learning, it's possible to detect signs of cognitive decline in driving behaviour. If these signs are identified, sharing this information with policyholders and providing driving advice can help reduce traffic accidents involving the elderly and extend their driving lifespan.


  • Regarding the data, damage assessment experts manually labeled accident videos obtained from dashcams, distinguishing those likely caused by cognitive decline from those that were not.

  • In this project, we utilized a vast array of telematics data, including accident data, to jointly develop an algorithm with Mind Foundry. This algorithm identifies policyholders who exhibit driving behaviors similar to those who have had accidents believed to be caused by cognitive decline.

  • In designing the algorithm, we considered the need to relay information back to the policyholders and thus employed explainable features.

Next Step

  • The algorithm is currently in the validation phase. Moving forward, we plan to collaborate with universities and medical institutions to enhance its accuracy.

  • This issue is of global relevance, and we are considering international expansion. As a first step, Aioi Nissay Dowa Europe has commenced research with the support of a grant from the UK government.

Additional Resources

Learn more about our project

bottom of page