| CPC B60W 30/0956 (2013.01) [A61B 5/02405 (2013.01); A61B 5/352 (2021.01); A61B 5/7257 (2013.01); B60W 40/08 (2013.01); B60W 50/0097 (2013.01); B60W 50/14 (2013.01); G06N 20/00 (2019.01); B60W 2050/146 (2013.01); B60W 2540/221 (2020.02); B60W 2556/10 (2020.02); B60W 2556/50 (2020.02)] | 35 Claims |
|
1. An operation support method configured to support operation of a vehicle by a computer including a processor and a memory, the method comprising:
obtaining, by the computer, first in-vehicle sensor data indicating a traveling state of the vehicle, the first in-vehicle sensor data being collected in a prior period, and hazard occurrence data having information on hazard occurrence from the first in-vehicle sensor data; generating, by the computer, an accident risk definition model configured to estimate a probability of the hazard occurrence by machine learning as an accident risk; receiving, by the computer, second in-vehicle sensor data indicating the traveling state of the vehicle, the second in-vehicle sensor data being collected in the prior period, to the accident risk definition model; generating, by the computer, accident risk estimation data by estimating the probability of the hazard occurrence;
receiving, by the computer, first biological index data and the accident risk estimation data, wherein the first biological index data is calculated in advance from first biological sensor data of a driver when the second in-vehicle sensor data is collected;
generating, by the computer, an accident risk prediction model by the machine learning, wherein the accident risk prediction model predicts the accident risk after a predetermined time;
acquiring, by the computer, second biological sensor data of the driver who is driving the vehicle;
calculating, by the computer, second biological index data indicating a state of the driver from the second biological sensor data; and
predicting, by the computer, the accident risk after the predetermined time by inputting the second biological index data to the accident risk prediction model,
wherein the acquiring includes:
acquiring heart rate data of the driver as the second biological sensor data,
calculating an RRI from the heart rate data and generating heart rate variability time-series data,
performing frequency spectral analysis of the heart rate variability time-series data, and
calculating, from a result of the frequency spectral analysis, a sum of an intensity of a low-frequency component of power spectral density and an intensity of a high-frequency component thereof as an autonomic nerve total power, and using the autonomic nerve total power as the biological index data.
|