Granted Patent B2

US 11,282,016 · App. 15/602,313 · Granted Mar 22, 2022

Inventors: Dima Kislovskiy (Pittsburgh, PA); David McAllister Bradley (Pittsburgh, PA)

Assignee: UATC, LLC

G06Q10/0635G06Q10/06315G08G1/202

Abstract

Claims (51)

1. An on-demand transportation management system comprising:

one or more processors; and

one or more memory resources storing instructions that, when executed by the one or more processors, cause the on-demand transportation management system to:

receive a transport request from a requesting user for a transportation service for a given region, the transport request indicating a pick-up location and a destination;

determine a candidate set of vehicles, within a proximity of the pick-up location, to service the transport request, wherein the candidate set of vehicles comprises an autonomous vehicle (AV);

receive, from one or more vehicles in the candidate set of vehicles, live AV data indicating a degradation level of the AV, wherein the degradation level defines at least one of sensor data quality of the AV, hardware performance of the AV, and software performance of the AV;

determine, based at least in part on the live AV data received from the one or more vehicles in the candidate set of vehicles, an individual risk value for each vehicle in the candidate set of vehicles for servicing the transport request using a machine-learned risk regressor, wherein the machine-learned risk regressor is trained using vehicle log data associated with the given region;

based, at least in part, on the individual risk value for each vehicle of the candidate set of vehicles, select a vehicle from the candidate set of vehicles to service the transport request; and

transmit, to the vehicle, instructions associated with performing the requested transportation service.

2. The on-demand transportation management system of claim 1 , wherein the vehicle log data comprises the live AV data.

3. The on-demand transportation management system of claim 1 , wherein the candidate set of vehicles comprises at least one of a safety-driven AV (SDAV) or a fully autonomous vehicle (FAV).

4. The on-demand transportation management system of claim 1 , wherein the executed instructions further cause the on-demand transportation management system to:

determine a set of routes between the pick-up location and the destination;

determine a generalized risk value for each route of the set of routes; and

select an optimal route from the set of routes based, at least in part, on the generalized risk value for each route.

5. The on-demand transportation management system of claim 4 , wherein the executed instructions cause the on-demand transportation management system to determine the generalized risk value for each route by aggregating individual risk values for each route segment of the route.

6. The on-demand transportation management system of claim 5 , wherein the executed instructions further cause the on-demand transportation management system to:

collect historical data identifying failed rides for the transportation service, each failed ride corresponding to a determined route;

wherein the executed instructions further cause the on-demand transportation management system to determine the generalized risk value for each route by determining a failure probability for the route based on the historical data across the route.

7. The on-demand transportation management system of claim 6 , wherein the generalized risk value for the route comprises a weighted sum between the failure probability and the aggregated individual risk values of respective route segments for the route.

8. The on-demand transportation management system of claim 5 , wherein the executed instructions further cause the on-demand transportation management system to:

determine a set of current conditions comprising at least one of weather conditions, road conditions, lighting conditions, time of day, or traffic conditions;

wherein the executed instructions cause the on-demand transportation management system to adjust the individual risk values for each route segment based on the set of current conditions; and

wherein the executed instructions cause the on-demand transportation management system to further determine a risk value for each driver in a candidate set of drivers based on the set of current conditions.

9. The on-demand transportation management system of claim 1 , wherein the candidate set of vehicles comprises one or more human-driven vehicles (HDVs), and wherein the executed instructions further cause the on-demand transportation management system to:

maintain a set of driver logs for drivers of the transportation service, each driver log indicating driving characteristics of a particular driver;

wherein the executed instructions cause the on-demand transportation management system to determine the individual risk value of each HDV in the candidate set of vehicles based on the driving characteristics of the driver of the HDV.

10. The on-demand transportation management system of claim 1 , wherein the executed instructions further cause the on-demand transportation management system to:

maintain a set of provider logs for each service provider of the transportation service, each provider log indicating provider characteristics of a corresponding service provider of the transportation service.

11. The on-demand transportation management system of claim 10 , wherein the provider logs comprise (i) vehicle logs for safety-driven AVs (SDAVs) and fully autonomous vehicles (FAVs) of the transportation service, the vehicle logs indicating a degradation level for each SDAV and FAV, and (ii) driver logs for drivers of human-driven vehicles (HDVs) of the transportation service, the driver logs indicating driving characteristics of the drivers.

12. The on-demand transportation management system of claim 11 , wherein the candidate set of vehicles comprises at least one safety-driven AV (SDAV), at least one fully autonomous vehicle (FAV), and at least one human-driven vehicle (HDV).

13. The on-demand transportation management system of claim 12 , wherein the executed instructions cause the on-demand transportation management system to determine the individual risk value of each SDAV, FAV, and driver of an HDV in the candidate set of vehicles based on the degradation level of each SDAV and FAV, and the driving characteristics of each driver of an HDV, in the candidate set of vehicles.

14. The on-demand transportation management system of claim 1 , wherein the executed instructions further cause the on-demand transportation management system to:

monitor progress of the selected vehicle along an optimal route;

monitor current conditions for the selected vehicle as the vehicle operates along the optimal route;

based at least in part on the current conditions, dynamically compute a trip remainder risk value for a remainder of the optimal route; and

based on the trip remainder risk value exceeding one or more risk thresholds for the selected vehicle, determine (i) an alternate set of routes from a current location of the selected vehicle to the destination, and (ii) determine a risk value for each of the alternate set of routes for the selected vehicle.

15. A non-transitory computer readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:

receive a transport request from a requesting user for a transportation service for a given region, the transport request indicating a pick-up location and a destination;

determine a candidate set of vehicles, within a proximity of the pick-up location, to service the transport request, wherein the candidate set of vehicles comprises an autonomous vehicle (AV);

receive, from one or more vehicles in the candidate set of vehicles, live AV data indicating a degradation level of the AV, wherein the degradation level defines at least one of sensor data quality of the AV, hardware performance of the AV, and software performance of the AV;

determine, based at least in part on the live AV data received from the one or more vehicles in the candidate set of vehicles, an individual risk value for each vehicle in the candidate set of vehicles for servicing the transport request using a machine-learned risk regressor, wherein the machine-learned risk regressor is trained using vehicle log data associated with the given region;

based, at least in part, on the individual risk value for each vehicle of the candidate set of vehicles, select a vehicle from the candidate set of vehicles to service the transport request; and

transmit, to the vehicle, instructions associated with performing the requested transportation service.

16. A computer-implemented method, the method being performed by one or more processors and comprising:

receiving a transport request from a requesting user for a transportation service for a given region, the transport request indicating a pick-up location and a destination;

determining a candidate set of vehicles, within a proximity of the pick-up location, to service the transport request, wherein the candidate set of vehicles comprises an autonomous vehicle (AV);

receiving, from one or more vehicles in the candidate set of vehicles, live AV data indicating a degradation level of the AV, wherein the degradation level defines at least one of sensor data quality of the AV, hardware performance of the AV, and software performance of the AV;

determining, based at least in part on the live AV data received from the one or more vehicles in the candidate set of vehicles, an individual risk value for each vehicle in the candidate set of vehicles for servicing the transport request using a machine-learned risk regressor, wherein the machine-learned risk regressor is trained using vehicle log data associated with the given region;

based, at least in part, on the individual risk value for each vehicle of the candidate set of vehicles, selecting a vehicle from the candidate set of vehicles to service the transport request; and

transmitting, to the vehicle, instructions associated with performing the requested transportation service.

Assignments (7)