Basic study of the estimation of the cognitive level using senior driver's driving behaviors

Japan is one of the leading country of the aging society. For traffic as well, the number of the traffic accidents caused by senior driver is increasing year by year. One of the reason is, number of the senior drivers who have cognitive impairment getting increasing because of length life-span. Currently, Japanese government moved into actions for example to have the campaign for returning of the senior drivers' licenses, or oblige senior driver who doesn't want to return his/her license to take a driving lessons at the driving school each 3 years. But to resolve this problem, society have to take a radically-new way to support senior drivers they can drive much more safely. It would be help for the senior people to keep healthy life. We thought, if we can find the driver who seems to have cognitive impairment, we should provide some advice and prevention training to retard progression. For this motivation, we consider the way to estimate drivers' cognition level from driving behavior. In this paper, we created estimation model to estimate cognition level. We analyzed and compared 654 senior drivers' driving data and cognition test result for the estimation. From driving data, mainly we extract 3 parameters which have significant difference between unimpaired senior driver and slightly impaired. Estimation was performed by RUS-Boost and study data were the 3 parameters. The estimation result was 77.2%. It seemed good result to recognize. It shows, there is possibility to estimate driver's cognition level by driving behavior then if impaired driver is found, some support can be adopted in early stage. It would become important strategy to make health longer and keep safely senior drivers' driving.