US 12,326,881 B2
Information processing method
Hideaki Bunazawa, Nagoya (JP); Keisuke Ninomiya, Nisshin (JP); Kunio Suzuki, Tokyo (JP); and Yoshitaka Tomiyama, Tokyo (JP)
Assigned to TOYOTA JIDOSHA KABUSHIKI KAISHA, Toyota (JP)
Filed by TOYOTA JIDOSHA KABUSHIKI KAISHA, Toyota (JP)
Filed on Aug. 27, 2024, as Appl. No. 18/815,857.
Claims priority of application No. 2023-165679 (JP), filed on Sep. 27, 2023.
Prior Publication US 2025/0103617 A1, Mar. 27, 2025
Int. Cl. G06F 16/28 (2019.01); B60W 40/105 (2012.01); B60W 40/107 (2012.01)
CPC G06F 16/285 (2019.01) [B60W 40/105 (2013.01); B60W 40/107 (2013.01); B60W 2556/10 (2020.02)] 5 Claims
OG exemplary drawing
 
1. An information processing method executed by information processing circuitry, the information processing method comprising:
collecting original data over a predetermined sensing period using sensors installed on a vehicle; and
obtaining extracted data by extracting data segments from the original data to reduce an amount of data used for analysis, wherein
the obtaining the extracted data includes:
calculating a first relative frequency distribution in the original data for each of feature amounts, the feature amounts being included in the original data, and the first relative frequency distribution being a relative frequency distribution for each of the feature amounts in the original data;
determining a setting of time windows, the setting of the time windows being determined to cut out the data segments corresponding to part of the predetermined sensing period from the original data, and the setting of the time windows being determined such that a total period of the time windows is shorter than the predetermined sensing period;
cutting out the data segments from the original data using the time windows;
obtaining the extracted data by combining all of the data segments that have been cut out using the time windows;
calculating a second relative frequency distribution for each of the feature amounts, the second relative frequency distribution being a relative frequency distribution in the extracted data;
calculating an error between the first relative frequency distribution and the second relative frequency distribution;
after obtaining the first relative frequency distribution, repeatedly executing a trial from the determining the setting of the time windows to the calculating the error while changing the setting of the time windows; and
selecting and outputting one or more of the settings of the time windows in which the error is less than a threshold value.