METHOD AND APPRATUS FOR DETECTING ANOMALY OF VEHICLE BASED ON EUCLIDEAN DISTANCE MEASURE

摘要

Disclosed are a method and a device for detecting abnormality of a vehicle based on a euclidean distance measurement technique, capable of rapidly and effectively detecting an abnormal vehicle state. The method for detecting abnormality of a vehicle based on the euclidean distance measurement technique includes: a step of being executed by a vehicle abnormality detecting device and collecting vehicle state data on an n (n is a natural number of 1 or greater) number of vehicles (k) in a predetermined section; a step of filtering the data in which there is no change of a data value among the vehicle state data and generating the filtered vehicle state data; a step of extracting an m (m is the natural number of 2 or greater) number of details in which a dimension is reduced by factor analysis of the filtered vehicle state data and generating the state data on each of the details; a step of extracting a reference value for detecting the abnormality of the vehicle by using traffic situation information; a step of calculating a euclidean distance (D_k_, called ′a first distance′) from the state data on a whole of the details of a vehicle to be analyzed and the reference value; a step of detecting the abnormality by using the first distance; a step of calculating a detail-based euclidean distance (called ′a second distance′) from the state data on each of the details of the vehicle to be analyzed and the reference value; a step of detecting the abnormality on the each of the details by using the second distance; and a step of providing an abnormality detection alarm including information on a cause of the abnormality of the vehicle when the abnormality of the vehicle is detected.