Abnormality Detection Using CAN Data

While there are high expectations for the utilization of automated driving, traffic accidents caused by driver error, such as stepping on the gas pedal and brake pedals by accident, remain a social problem. The goal of this study is to detect driver errors and abnormalities in the vehicle body using CAN data, which are sensor data obtained from inside the vehicle and information from in-vehicle cameras. This is a collaborative study with Toyota Motor Corporation, and the experiments were conducted using actual data from abnormalities and accidents that were not included in the open data set.


Search this site

Recent News