Method and apparatus for driver identification leveraging telematics data

A computer program product for identifying a driver of a vehicle based on recorded telematics data, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions being executable by a processor to cause the processor to perform a method comprising:

• receiving, by a trip split engine, a sequence of measurement values from one or more telematics sensors measuring features relating to the movement of a vehicle, wherein the sequence of measurement values represents a trip segment of an arbitrary period of time for the vehicle;
• defining, by a feature map transformer, a number of instantaneous dimensions for measuring driver identification;
• for each dimension, calculating by the feature map transformer, statistical features within a given time frame;
• building, by the feature map transformer, a feature map, in which map a first dimension represents the time frame and another dimension represents the calculated statistical features;
• extracting, by a deep neural network model engine, a set of driving features from the feature map using a deep neural network model;
• comparing, by a similarity model engine, the set of extracted driving features with previously extracted sets of driving features to create a set of similarity metrics between two drivers;
• training, by a trip classification engine, a multi-variable classification model based on the similarity metrics;
• and
• using, by the trip classification engine, the classification model and the similarity metrics to determine whether data pertaining to a new trip segment should be associated with an unknown driver or with a known driver, wherein the determination is made by referencing recorded telematics data.

A system comprising for identifying a driver of a vehicle based on recorded telematics data:

• a trip split engine, configured to:
• receive a sequence of measurement values from one or more telematics sensors measuring features relating to the movement of a vehicle, wherein the sequence of measurement values represents a trip segment of an arbitrary period of time for the vehicle;
• a feature map transformer configured to:
• define a number of instantaneous dimensions for measuring driver identification, for each dimension, calculate statistical features within a given time frame, and
• build a feature map, in which map a first dimension represents the time frame and another dimension represents the calculated statistical features;
• a deep neural network model engine configured to:
• extract a set of driving features from the feature map using a deep neural network model;
• a similarity model engine configured to:
• compare the set of extracted driving features with previously extracted sets of driving features to create a set of similarity metrics between two drivers;
• and
• a trip classification/clustering engine configured to:
• train a multi-variable classification model based on the similarity metrics, and
• use the classification model and the similarity metrics to determine whether data pertaining to a new trip segment should be associated with an unknown driver or with a known driver, wherein the determination is made by referencing recorded telematics data.