Detecting service provider state based on received and historic telematics data and service provider characteristics
View Patent ↗A network system is configured to facilitate a safe service coordination environment. The network system detects abnormal user states (e.g., sleepiness, inebriation, etc.) and provides corrective recommendations to those users. To do so, client devices executing an application send information associated with the service to the network system. The network system determines an aggregate set of service characteristics based on the aggregate telematics data. The network system also determines a provider profile for a provider providing service in the environment describing the deviation of the user's driving characteristics from the aggregate service characteristics. Based on the difference between the aggregate characteristics and the user profile, the network system determines the provider state using a state verification process. Based on the determined provider state, the network system determines a set of corrective recommendations and sends the corrective recommendations to the provider device.
1. A method for determining abnormal driving characteristics comprising:
receiving current telematics data from a client device for a provider providing a service along at least a portion of a route, the telematics data describing movement of the client device during the service;
determining a typical frequency of one or more abnormal behaviors associated with the portion of the route based on a set of aggregate telematics data gathered from a set of client devices;
generating a provider profile with identifying information of the provider, including a deviation from the typical frequency of the one or more abnormal behaviors for the provider;
determining an expected frequency of the one or more abnormal behaviors for the provider along the portion of the route by adjusting the typical frequency of the one or more abnormal behaviors based on the provider profile;
generating an abnormality score reflecting a difference between a current frequency of the one or more abnormal behaviors of the current telematics data and the expected frequency of the one or more abnormal behaviors,
when the abnormality score exceeds a threshold, identifying at least one abnormal driving characteristic of the provider;
generating a corrective recommendation based on the at least one abnormal driving characteristic of the provider.
2. The method of claim 1 , wherein
the provider provides a service along a route and the received telematics data is associated with the route;
the aggregate telematics data of the set of client devices is associated with the route; and
determining the abnormality score is based on the route.
3. The method of claim 2 , wherein determining the abnormality score is based on the route comprises:
determining a road segment of the route based on the received telematics data; and
comparing the aggregate telematics data of the set of client devices on the road segment to the telematics data of the client device on the road segment.
4. The method of claim 2 , wherein determining the abnormality score based on the route accounts for real-time deviations of the provider from the route.
5. The method of claim 1 , wherein
the provider profile further specifies an average frequency of the one or more abnormal behaviors and a standard deviation from the average frequency of the one or more abnormal behaviors; and
the abnormality score reflects the standard deviation from the average frequency of the one or more abnormal behaviors.
6. The method of claim 1 , wherein the threshold is a standard deviation of three sigmas away from the average frequency of the one or more abnormal behaviors.
7. The method of claim 1 , comprising:
creating the provider profile based on the normal behavior of the set of client devices.
8. The method of claim 7 , wherein the created provider profile is based on any of a country, a state, or a city in which the provider and the set of client devices is providing the service.
9. The method of claim 7 , comprising:
modifying the created provider profile based on the received telematics data of the provider as the provider provides the service.
10. The method of claim 1 , wherein identifying unusual driving characteristics of the provider is in response to a trigger, the trigger being one or more of: an input from a requesting user associated with the provider using the service, a threshold amount of time the provider has been providing the service, a variation in the abnormality score, or a variation in the telematics data.
11. The method of claim 1 , wherein the abnormality score is associated with one or more of: a speed of the provider, a change in the speed of the provider, a set of previously identified unusual driving characteristics, a position of the provider, a set of received interactions from the provider using the client device, a route of the service, or a requesting user of the service associated with the provider.
12. The method of claim 1 , comprising:
determining the state of the service provider, the state associated with the unusual driving characteristics of the provider; and
wherein the state of the service provider includes any of: inebriation, sleep deprivation, or mental state.
13. The method of claim 12 , wherein the corrective recommendation is based on the state of the provider.
14. The method of claim 1 , comprising:
providing the corrective recommendation to the provider; and
wherein the corrective recommendation determines the cause of unusual driving characteristics.
15. The method of claim 1 , wherein the corrective recommendation is an interactive test for the provider, the interactive test determining a subsequent corrective recommendation for the provider.
16. The method of claim 1 , wherein the corrective recommendation is based on at least one previous corrective recommendation associated with the provider.
17. A non-transitory computer-readable storage medium storing computer program instructions executable by a processor of a system to perform steps for determining abnormal driving characteristics comprising:
receiving current telematics data from a client device for a provider providing a service along at least a portion of a route, the telematics data describing movement of the client device during the service;
determining a typical frequency of one or more abnormal behaviors associated with the portion of the route based on a set of aggregate telematics data gathered from a set of client devices;
generating a provider profile with identifying information of the provider, including a deviation from the typical frequency of the one or more abnormal behaviors for the provider;
determining an expected frequency of the one or more abnormal behaviors for the provider along the portion of the route by adjusting the typical frequency of the one or more abnormal behaviors based on the provider profile;
generating an abnormality score reflecting a difference between a current frequency of the one or more abnormal behaviors of the current telematics data and the expected frequency of the one or more abnormal behaviors,
when the abnormality score exceeds a threshold, identifying at least one abnormal driving characteristic of the provider;
generating a corrective recommendation based on the at least one abnormal driving characteristic of the provider.
18. The non-transitory computer-readable storage medium of claim 17 , wherein
the provider provides a service along a route and the received telematics data is associated with the route;
the behavior of a set of client devices is associated with the route; and
determining the abnormality score is based on the route.
19. The non-transitory computer-readable storage medium of claim 17 , wherein determining the abnormality score is based on the route further comprises:
determining a road segment of the route based on the received telematics data; and
comparing the aggregate telematics data of the set of client devices providing the service on the road segment to the telematics data of the client device on the road segment.
20. A method for remediating abnormal driving characteristics, the method comprising:
generating, for a first provider of a transportation service, a provider profile, the provider profile including a deviation of the first provider's driving characteristics from a typical frequency of abnormal behavior;
determining, for a portion of a route, a typical frequency of abnormal behaviors along the portion of the route, the typical frequency based on historical telematics data collected from a plurality of client devices of other providers traveling along the portion of the route;
determining an expected frequency of the one or more abnormal behaviors for the first provider along the portion of the route by adjusting the typical frequency of the one or more abnormal behaviors along the portion of the route based on the deviation included in the first provider profile;
receiving current telematics data from a client device of the first provider while the first provider is providing a transportation service, the current telematics data describing movement of the client device along the portion of the route;
generating an abnormality score reflecting a difference between a current frequency of the one or more abnormal behaviors observed in the received current telematics data and the expected frequency of the one or more abnormal behaviors;
responsive to the abnormality score exceeding a threshold, identifying at least one abnormal driving characteristic of the provider; and
providing a corrective recommendation based on the at least one abnormal driving characteristic of the first provider.