CLOUD-BASED VEHICULAR TELEMATICS SYSTEMS AND METHODS FOR GENERATING HYBRID EPOCH DRIVER PREDICTIONS AND HAZARD ALERTS

1 . A computer-implemented method for generating a hazard alert for a user and implemented by a computing device, the method comprising:

for each first user of a group of first users:

receiving, from a mobile device associated with a first user, prior telematics data indicative of an operation of a vehicle by the first user during one or more prior trips in a prior epoch;

receiving, from one or more sensors of the mobile device, raw sensor data;

transforming the raw sensor data into recent telematics data indicative of characteristics associated with the operation of the vehicle by the first user during one or more recent trips in a current epoch, wherein the recent telematics data are compressed to have a different dimension than the raw sensor data;

generating a prior epoch score based at least in part upon the prior telematics data;

wherein:

the current epoch is associated with a plurality of time segments;

each recent trip of the one or more recent trips is associated with one or more time segments of the plurality of time segments;

assigning the one or more time segments of the plurality of time segments with the prior epoch score;

predicting, for each recent trip of the one or more recent trips, a driving metric associated with at least one of braking or speeding by applying a trained machine learning model to the recent telematics data, wherein the trained machine learning model is trained using one or more historical epoch scores and one or more historical driving metrics, wherein the one or more historical epoch scores and the one or more historical driving metrics are extracted from historical telematics data;

generating, for each recent trip of the one or more recent trips, a partial current epoch score based at least in part upon the driving metric;

updating, for each recent trip of the one or more recent trips, the one or more time segments with the partial current epoch score; and

generating a hybrid current epoch score based at least in part upon the prior epoch score and the partial current epoch score;

determining a number of users of the group of first users whose one or more behavioral changes exceeds or equals to one or more pre-determined alert thresholds based at least in part upon the hybrid current epoch score; and

generating, upon determining that the number of users is greater than or equal to a pre-determined trend-threshold, one or more hazard alerts for a group of second users.

2 . The computer-implemented method of claim 1 , wherein:

the prior epoch is associated with a prior epoch time;

the one or more recent trips is associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch; and

the generating the hybrid current epoch score includes:

determining a remaining time of the current epoch as the prior epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the prior epoch score multiplied by the remaining time; and

the partial current epoch score multiplied by the recent driving time.

3 . The computer-implemented method of claim 1 , wherein:

the generating the hybrid current epoch score includes generating the hybrid current epoch score based at least in part upon:

the partial current epoch score for each time segment of the plurality of time segments that was updated; and

the prior epoch score for each time segment of the plurality of time segments that was not updated.

4 . The computer-implemented method of claim 1 , further comprising for each first user of the group of first users:

receiving, from the mobile device, historic telematics data indicative of the operation of the vehicle by the first user during one or more historic trips in one or more historic epochs; and

determining a historic epoch time based at least in part upon the historic telematics data, the historic epoch time being an average total time of a historic epoch of the one or more historic epochs;

wherein:

the one or more recent trips are associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch;

the generating the hybrid current epoch score includes:

determining a remaining time of the current epoch as the historic epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the partial current epoch score multiplied by a ratio of the elapsed time to a time historically driven per epoch; and

the prior epoch score multiplied by a ratio of the remaining time to the time historically driven per epoch.

5 . The computer-implemented method of claim 1 , wherein:

the prior telematics data include:

prior qualitative data indicative of one or more prior driving behaviors of the first user during the one or more prior trips; and

prior quantitative data indicative of a prior driving time in the prior epoch; the recent telematics data include:

recent qualitative data indicative of one or more recent driving behaviors of the first user during the one or more recent trips; and

recent quantitative data indicative of a recent driving time in the current epoch;

the generating the prior epoch score includes generating the prior epoch score based at least in part upon the prior qualitative data and the prior quantitative data;

the generating the partial current epoch score includes generating the partial current epoch score based at least in part upon the recent qualitative data and the recent quantitative data; and

the generating the hybrid current epoch score includes generating the hybrid current epoch score based at least in part upon the prior qualitative data, the prior quantitative data, the recent qualitative data, and the recent quantitative data.

6 . The computer-implemented method of claim 1 , wherein the:

receiving the recent telematics data includes receiving the recent telematics data at least one of in real-time, in near real-time, after completion of each recent trip of the one or more recent trips, continuously, and at a pre-determined time;

generating the partial current epoch score includes generating the partial current epoch score at least one of in real-time, in near real-time, after completion of each recent trip of the one or more recent trips, continuously, and at a pre-determined time; and

generating the hybrid current epoch score includes generating the hybrid current epoch score at least one of in real-time, in near real-time, after completion of each recent trip of the one or more recent trips, continuously, and at a pre-determined time.

7 . The computer-implemented method of claim 1 , further comprising:

receiving third-party information associated with at least one of the one or more prior trips and the one or more recent trips;

wherein at least one of the:

generating the prior epoch score includes generating the prior epoch score based at least in part upon the prior telematics data and the third-party information; and

generating the partial current epoch score includes generating the partial current epoch score based at least in part upon the recent telematics data and the third-party information.

8 . The computer-implemented method of claim 1 , further comprising:

presenting the one or more hazard alerts to the group of second users.

9 . The computer-implemented method of claim 1 , wherein the one or more hazard alerts is indicative of one or more driving conditions associated with one or more locations.

10 . The computer-implemented method of claim 9 , wherein the one or more locations includes one of a section of a highway, an intersection, an entrance ramp, and an exit ramp.

11 . The computer-implemented method of claim 1 , further comprising:

for each first user of the group of first users:

detecting, using the one or more sensors associated with the mobile device, whether the vehicle is being operated by the first user; and

collecting, in response to detecting that the vehicle is being operated by the first user, telematics data using the one or more sensors and based upon one or more triggering events comprising at least one of the one or more sensors exceeding a threshold amount of sensor measurements, wherein the recent telematics data are indicative of the operation of the vehicle by the first user during the one or more recent trips.

12 . The computer-implemented method of claim 1 , wherein at least one of the:

generating the prior epoch score includes generating, using a neural network, the prior epoch score based at least in part upon the prior telematics data;

generating the partial current epoch score includes generating, using a neural network, the partial current epoch score based at least in part upon the recent telematics data; and

generating the hybrid current epoch score includes generating, using a neural network, the hybrid current epoch score based at least in part upon the prior epoch score and the partial current epoch score.

13 . A system for generating a hazard alert, the system comprising:

a data receiving module configured to:

for each first user of a group of first users:

receive, from at least one of a mobile device associated with a first user, prior telematics data indicative of the operation of a vehicle by the first user during one or more prior trips in a prior epoch; and

receive, from one or more sensors of the mobile device, raw sensor data;

transform the raw sensor data into recent telematics data indicative of characteristics associated with the operation of the vehicle by the first user during one or more recent trips in a current epoch, wherein the recent telematics data are compressed to have a different dimension than the raw sensor data;

wherein:

the current epoch is associated with a plurality of time segments;

each recent trip of the one or more recent trips is associated with one or more time segments of the plurality of time segments;

an epoch score generating module configured to:

for each first group user of a plurality of first group users:

generate a prior epoch score based at least in part upon the prior telematics data;

assign the one or more time segments of the plurality of time segments with the prior epoch score;

predict, for each recent trip of the one or more recent trips, a driving metric associated with at least one of braking or speeding by applying a trained machine learning model to the recent telematics data, wherein the trained machine learning model is trained using one or more historical epoch scores and one or more historical driving metrics, wherein the one or more historical epoch scores and the one or more historical driving metrics are extracted from historical telematics data;

generate, for each recent trip of the one or more recent trips, a partial current epoch score based at least in part upon the driving metric;

update, for each recent trip of the one or more recent trips, the one or more time segments with the partial current epoch score; and

generate a hybrid current epoch score based at least in part upon the prior epoch score and the partial current epoch score; and

an alert module configured to:

determine a number of users of the group of first users whose one or more behavioral changes exceeds or equals to one or more pre-determined alert thresholds based at least in part upon the hybrid current epoch score; and

generate, upon determining that the number of users is greater than or equal to a pre-determined trend-threshold, one or more hazard alerts for a group of second users.

14 . The system of claim 13 , wherein:

the prior epoch is associated with a prior epoch time;

the one or more recent trips is associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch; and

the epoch score generating module is configured to generate the hybrid current epoch score by at least:

determining a remaining time of the current epoch as the prior epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the prior epoch score multiplied by the remaining time; and

the partial current epoch score multiplied by the recent driving time.

15 . The system of claim 13 , wherein:

the epoch score generating module is further configured to:

generate the hybrid current epoch score based at least in part upon:

the partial current epoch score for each time segment of the plurality of time segments that was updated; and

the prior epoch score for each time segment of the plurality of time segments that was not updated.

16 . The system of claim 13 , wherein:

the data receiving module is further configured to:

for each first user of the group of first users:

receive, from the mobile device, historic telematics data indicative of the operation of the vehicle by the first user during one or more historic trips in one or more historic epochs;

the one or more recent trips is associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch; and

the epoch score generating module is configured to:

determine a historic epoch time based at least in part upon the historic telematics data, the historic epoch time being an average total time of a historic epoch of the one or more historic epochs;

generate the hybrid current epoch score by at least:

determining a remaining time of the current epoch as the historic epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the partial current epoch score multiplied by a ratio of the elapsed time to a time historically driven per epoch; and

the prior epoch score multiplied by a ratio of the remaining time to the time historically driven per epoch.

17 . A non-transitory computer-readable medium with instructions stored thereon, that upon execution by a processor, causes the processor to perform:

for each first user of a group of first users:

receiving, from a mobile device associated with a first user, prior telematics data indicative of the operation of a vehicle by the first user during one or more prior trips in a prior epoch;

receiving, from one or more sensors of the mobile device, raw sensor data;

transforming the raw sensor data into recent telematics data indicative of characteristics associated with the operation of the vehicle by the first user during one or more recent trips in a current epoch, wherein the recent telematics data are compressed to have a different dimension than the raw sensor data;

generating a prior epoch score based at least in part upon the prior telematics data; wherein:

the current epoch is associated with a plurality of time segments;

each recent trip of the one or more recent trips is associated with one or more time segments of the plurality of time segments;

assigning the one or more time segments of the plurality of time segments with the prior epoch score;

predicting, for each recent trip of the one or more recent trips, a driving metric associated with at least one of braking or speeding by applying a trained machine learning model to the recent telematics data, wherein the trained machine learning model is trained using one or more historical epoch scores and one or more historical driving metrics, wherein the one or more historical epoch scores and the one or more historical driving metrics are extracted from historical telematics data;

generating, for each recent trip of the one or more recent trips, a partial current epoch score based at least in part upon the driving metric;

updating, for each recent trip of the one or more recent trips, the one or more time segments with the partial current epoch score; and

generating a hybrid current epoch score based at least in part upon the prior epoch score and the partial current epoch score;

determining a number of users of the group of first users whose one or more behavioral changes exceeds or equals to one or more pre-determined alert thresholds based at least in part upon the hybrid current epoch score; and

generating, upon determining that the number of users is greater than or equal to a pre-determined trend-threshold, one or more hazard alerts for a group of second users.

18 . The non-transitory computer-readable medium of claim 17 , wherein:

the prior epoch is associated with a prior epoch time;

the one or more recent trips is associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch; and

the generating the hybrid current epoch score includes:

determining a remaining time of the current epoch as the prior epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the prior epoch score multiplied by the remaining time; and

the partial current epoch score multiplied by the recent driving time.

19 . The non-transitory computer-readable medium of claim 17 , wherein:

the generating the hybrid current epoch score includes generating the hybrid current epoch score based at least in part upon:

the partial current epoch score for each time segment of the plurality of time segments that was updated; and

the prior epoch score for each time segment of the plurality of time segments that was not updated.

20 . The non-transitory computer-readable medium of claim 17 , that upon execution of the processor, further causes the processor to perform:

for each first user of the group of first users:

receiving, from the mobile device, historic telematics data indicative of the operation of the vehicle by the first user during one or more historic trips in one or more historic epochs; and

determining a historic epoch time based at least in part upon the historic telematics data, the historic epoch time being an average total time of a historic epoch of the one or more historic epochs;

wherein:

the one or more recent trips is associated with a recent driving time and an elapsed time, the recent driving time being an amount of time the vehicle was under operation in the current epoch, the elapsed time being the amount of time that has passed in the current epoch;

the generating the hybrid current epoch score includes:

determining a remaining time of the current epoch as the historic epoch time minus the elapsed time; and

determining the hybrid current epoch score based at least in part upon adding:

the partial current epoch score multiplied by a ratio of the elapsed time to a time historically driven per epoch; and

the prior epoch score multiplied by a ratio of the remaining time to the time historically driven per epoch.