A Study On Finding The Optimal Time For Automatic Transition To Self-Driving Mode

Topic of self-driving mode and transition to this mode is one of the trend topics of biomedical engineering and artificial intelligence studies. Sleeplessness and sleep efficiency to cause inattention in driving and accidents. This study aimed to investigate convenient time to transit self-driving mode respect to number of accidents and sleep efficiency of driver by using Support Vector Machines and K-Nearest neighbors classification algorithms to reduce the accidents. Approximate entropy and Lyapunov exponent for Electroencephalography and dominant frequency, ratio of power of high frequency to low frequency, area under the curve and derivative respiration signals were extracted. This proposal method achieves 93.33% and 100% accuracies to classify drivers and transit car to self-driving mode respect to two criteria.