Analyzing telematics data within heterogeneous vehicle populations

• US 9,846,979 B1
• Filed: 06/16/2016
• Issued: 12/19/2017
• Est. Priority Date: 06/16/2016
• Status: Active Grant

• Current Assignee: Mojio Inc.
• Original Assignee: Mojio Inc.
• Inventors: Sainaney, Narayan, Vora, Tejas
• Primary Examiner(s): Butler, Rodney
• Application Number:
  US15/184,771
• Publication Number:
  US 20170365110A1
• Time in Patent Office: 551 Days
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance
• CPC Class Codes: G07C 5/002 Analysing tachograph charts arrangements for graph-reading in general G06K11/00
• CPC Class Codes: G07C 5/008 communicating information to a remotely located station transmission systems for measured values G08C
• CPC Class Codes: G07C 5/0808 Diagnosing performance data testing of vehicles G01M17/00; testing of electrical installation on vehicles G01R31/005
• CPC Class Codes: G07C 5/0825 using optical means
• CPC Class Codes: G07C 5/085 using electronic data carriers

Abstract

A computing system implements a telematics service that obtains a set of vehicle telematics data for each vehicle of a population. Within each set of vehicle telematics data, a set of time-based measurements for a measurement type is identified. The set of time-based measurements identified for each vehicle are combined to obtain a combined set of time-based observations for the measurement type across the population of vehicles or a sub-set of the population defined by vehicle make, model, and/or year of production. An outlier observation is identified from among the combined set of time-based observations. A determination is made whether the outlier observation is part of a temporary deviation or a persistent deviation. For a temporary deviation, an impact of the outlier observation on the set of time-based measurements is reduced. For a persistent deviation, the outlier observation is programmatically characterized.

Claims

1. A method performed by a computing system implementing a telematics service, the method comprising:

obtaining a set of vehicle telematics data for each vehicle of a population of vehicles,

wherein for each vehicle of the population, the set of vehicle telematics data is obtained by receiving communications transmitted from that vehicle by a vehicle-based telematics device via a wireless communications network;

identifying, within each set of vehicle telematics data, a vehicle identifier of the vehicle from which that set of vehicle telematics data originated and a set of time-based measurements for a measurement type indicated by a measurement type identifier contained within the set of vehicle telematics data, the set of time-based measurements captured by one or more sensors of the vehicle and received from the one or more sensors along with the vehicle identifier of the vehicle by the vehicle-based telematics device via a physical connector of an on-board diagnostics (OBD) interface;

combining the set of time-based measurements identified for each vehicle to obtain a combined set of time-based observations for the measurement type indicated by the measurement type identifier contained in each set of vehicle telematics data across the population of vehicles;

applying a statistical model to the combined set of time-based observations to obtain a set of one or more probability distributions;

identifying an outlier observation from among the combined set of time-based observations that is located outside of the set of one or more probability distributions;

for the outlier observation, identifying the vehicle identifier attributed to the outlier observation;

determining whether the outlier observation is part of a temporary deviation or a persistent deviation;

if the outlier observation is part of a temporary deviation, then reducing an impact of the outlier observation on the set of time-based measurements of which the outlier observation is a member to obtain an augmented set of time-based measurements;

if the outlier is part of a persistent deviation, then programmatically characterizing the outlier observation based, at least in part, on the set of time-based measurements of which the outlier observation is a member to obtain a characterization of the persistent deviation; and

outputting the augmented set of time-based measurements and/or the characterization of the persistent deviation by forwarding the augmented set of time-based measurements and/or the characterization of the persistent deviation to a mobile device of a subscriber associated with the vehicle identifier attributed to the outlier observation.

2. The method of claim 1, wherein the population of vehicles is a subset of a broader population of vehicles, and wherein the method further comprises:

obtaining a set of vehicle telematics data for each vehicle of the broader population of vehicles;

identifying one or more of a vehicle make, vehicle model, and/or vehicle year of production of each vehicle of the broader population of vehicles based on a vehicle identifier contained within each set of vehicle telematics data; and

defining the subset of the broader population as being limited to vehicles that share one or more of a similar vehicle make, a similar vehicle model, or a similar vehicle year of production with each other;

wherein the combined set of time-based observations for the measurement type is limited to the subset of the broader population.

3. The method of claim 2, wherein the subset of the broader population is limited to vehicles that share two or more of a similar vehicle make, a similar vehicle model, or a similar vehicle year of production with each other.

4. The method of claim 2, wherein the subset of the broader population is limited to vehicles that share three of a similar vehicle make, a similar vehicle model, or a similar vehicle year of production with each other.

5. The method of claim 2, wherein the subset of the broader population is limited to vehicles that share a similar vehicle make and a similar vehicle model with each other, and are within a predefined range of a vehicle year of production from each other.

6. The method of claim 1, further comprising:

generating the statistical model for the population of vehicles by defining the one or more probability distributions based on an attribute of the combined set of time-based observations.

7. The method of claim 1, wherein the characterization of the persistent deviation includes an indication that a defective or degraded component of a vehicle identified by the vehicle identifier attributed to the outlier observation is responsible for the persistent deviation.

8. The method of claim 1, wherein the characterization of the persistent deviation includes an indication that a servicing task for a vehicle identified by the vehicle identifier attributed to the outlier observation is due.

9. The method of claim 1, wherein reducing the impact of the outlier observation on the set of time-based measurements includes filtering the outlier observation from the set of time-based measurements.

10. The method of claim 1, wherein reducing the impact of the outlier observation on the set of time-based measurements includes modifying a value of the outlier observation to be within or closer to the set of one or more probability distributions.

11. The method of claim 1, further comprising:

if the outlier observation is part of a temporary deviation, assigning an operational score to the vehicle identifier based on the augmented set of time-based measurements, and outputting the operational score;

if the outlier is part of a persistent deviation, then assigning the operational score to the vehicle identifier based on the set of time-based measurements in which the operational score is output as the characterization of the persistent deviation.

12. The method of claim 1, further comprising:

assigning a relative score to the augmented set of time-based measurements that is based on a value of the augmented set of time-based measurements relative to the combined set of time-based measurements; and

wherein outputting the augmented set of time-based measurements includes outputting the relative score.

20. The method of claim 1, wherein determining whether the outlier observation is part of a temporary deviation or a persistent deviation is based on one or more factors including:

(1) a time-based duration of the deviation for the outlier observation for the vehicle, (2) a time-density or quantity of outlier observations within a period of time for the vehicle, and (3) a direction of the deviation from a normalized value or range of values.

13. A computing system including one or more computing devices, comprising:

a logic subsystem including one or more physical logic devices; and

a storage subsystem including one or more physical memory devices having instructions stored thereon executable by the one or more physical logic devices of the logic subsystem to;

obtain a set of vehicle telematics data for each vehicle of a population of vehicles,

wherein for each vehicle of the population, the set of vehicle telematics data is obtained by receiving communications transmitted from that vehicle by a vehicle-based telematics device via a wireless communications network;

identify, within each set of vehicle telematics data, a vehicle identifier of the vehicle from which that set of vehicle telematics data originated and a set of time-based measurements for a measurement type indicated by a measurement type identifier contained within the set of vehicle telematics data, the set of time-based measurements captured by one or more sensors of the vehicle and received from the one or more sensors along with the vehicle identifier of the vehicle by the vehicle-based telematics device via a physical connector of an on-board diagnostics (OBD) interface;

combine the set of time-based measurements identified for each vehicle to obtain a combined set of time-based observations for the measurement type indicated by the measurement type identifier contained in each set of vehicle telematics data across the population of vehicles;

apply a statistical model to the combined set of time-based observations to obtain a set of one or more probability distributions;

identify an outlier observation from among the combined set of time-based observations that is located outside of the set of one or more probability distributions;

for the outlier observation, identify the vehicle identifier attributed to the outlier observation;

determine whether the outlier observation is part of a temporary deviation or a persistent deviation;

if the outlier observation is part of a temporary deviation, then reduce an impact of the outlier observation on the set of time-based measurements of which the outlier observation is a member to obtain an augmented set of time-based measurements;

if the outlier is part of a persistent deviation, then programmatically characterize the outlier observation based, at least in part, on the set of time-based measurements of which the outlier observation is a member to obtain a characterization of the persistent deviation; and

output the augmented set of time-based measurements and/or the characterization of the persistent deviation by forwarding the augmented set of time-based measurements and/or the characterization of the persistent deviation to a mobile device of a subscriber associated with the vehicle identifier attributed to the outlier observation.

14. The computing system of claim 13, wherein the population of vehicles is a subset of a broader population of vehicles, and wherein the instructions are further executable by the one or more physical logic devices to:

obtain a set of vehicle telematics data for each vehicle of the broader population of vehicles;

identify one or more of a vehicle make, vehicle model, and/or vehicle year of production of each vehicle of the broader population of vehicles based on a vehicle identifier contained within each set of vehicle telematics data; and

define the subset of the broader population as being limited to vehicles that share one or more of a similar vehicle make, a similar vehicle model, or a similar vehicle year of production with each other;

wherein the combined set of time-based observations for the measurement type is limited to the subset of the broader population.

15. The computing system of claim 14, wherein the subset of the broader population is limited to vehicles that share a similar vehicle make and a similar vehicle model with each other, and are within a predefined range of a vehicle year of production from each other.

16. The computing system of claim 13, wherein the instructions are further executable by the one or more physical logic devices to:

generate the statistical model for the population of vehicles by defining the set of the one or more probability distributions based on an attribute of the combined set of time-based observations.

17. The computing system of claim 13, wherein the characterization of the persistent deviation includes:

an indication that a defective or degraded component of a vehicle identified by the vehicle identifier attributed to the outlier observation is responsible for the persistent deviation, or

an indication that a servicing task for a vehicle identified by the vehicle identifier attributed to the outlier observation is due.

18. The computing system of claim 13, wherein the impact of the outlier observation on the set of time-based measurements is reduced by:

filtering the outlier observation from the set of time-based measurements, or

modifying a value of the outlier observation to be within or closer to the set of one or more probability distributions.

19. The computing system of claim 13, wherein the outlier observation is determined to be part of a temporary deviation or a persistent deviation based on one or more factors including:

(1) a time-based duration of the deviation for the outlier observation for the vehicle, (2) a time-density or quantity of outlier observations within a period of time for the vehicle, and (3) a direction of the deviation from a normalized value or range of values.