Estimating trip duration based on vehicle reroute probabilities
The subject disclosure relates to ways to improve route duration calculations e.g., estimated time of arrival (ETA) approximations, by taking consideration of reroute probabilities along a given vehicle path. In some aspects, the disclosed technology encompasses a process including steps for identifying a route between a first destination to a second destination, determining a reroute likelihood associated with at least one AV maneuver along the route, and calculating an estimated time of arrival (ETA) based on the reroute likelihood associated with at least one AV maneuver. Systems and machine-readable media are also provided.
1. A computer-implemented method for determining an arrival time of an autonomous vehicle (AV), comprising:
identifying a route between a first map location and a second map location;
obtaining, from each respective autonomous vehicle (AV) of a set of AVs, sensor data indicating whether the respective AV successfully completed an AV maneuver at one or more locations along the route;
based on one or more capabilities of a particular AV and the sensor data indicating whether the respective AV successfully completed the AV maneuver at one or more locations along the route, determining a likelihood of a failure by the particular AV to perform one or more AV maneuvers at one or more portions of the route;
based on the likelihood of the failure by the particular AV to perform the one or more AV maneuvers at the one or more portions of the route, determining a reroute likelihood associated with at least one AV maneuver from the one or more AV maneuvers;
calculating an estimated time of arrival (ETA) based on the reroute likelihood associated with at least one AV maneuver; and
sending the calculated ETA to a computing device associated with a user.
2. The computer-implemented method of claim 1 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a map location.
3. The computer-implemented method of claim 1 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a lane identifier (ID).
4. The computer-implemented method of claim 1 , wherein calculating the ETA is further based on a time of day.
5. The computer-implemented method of claim 1 , wherein the one or more capabilities of the particular AV comprise an AV software version of the particular AV, and wherein the likelihood of the failure by the particular AV to perform the one or more AV maneuvers at the one or more portions of the route is at least partly based on the AV software version of the particular AV.
6. The computer-implemented method of claim 1 , wherein the reroute likelihood is based on at least one of an AV maneuver type and traffic conditions.
7. The computer-implemented method of claim 1 , further comprising:
updating one or more map priors following the at least one AV maneuver.
8. A system for determining an arrival time of an autonomous vehicle (AV), the system comprising:
one or more processors; and
a computer-readable medium comprising instructions stored therein, which when executed by the one or more processors, cause the one or more processors to perform operations comprising:
identifying a route between a first map location and a second map location;
obtaining, from each respective autonomous vehicle (AV) of a set of AVs, sensor data indicating whether the respective AV successfully completed an AV maneuver at one or more locations along the route;
based on one or more capabilities of a particular AV and the sensor data indicating whether the respective AV successfully completed the AV maneuver at one or more locations along the route, determining a likelihood of a failure by the particular AV to perform one or more AV maneuvers at one or more portions of the route;
based on the likelihood of the failure by the particular AV to perform the one or more AV maneuvers at the one or more portions of the route, determining a reroute likelihood associated with at least one AV maneuver from the one or more AV maneuvers;
calculating an estimated time of arrival (ETA) based on the reroute likelihood associated with at least one AV maneuver; and
sending the calculated ETA to a computing device associated with a user.
9. The system of claim 8 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a map location.
10. The system of claim 8 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a lane identifier (ID).
11. The system of claim 8 , wherein calculating the ETA is further based on a time of day.
12. The system of claim 8 , wherein the one or more capabilities of the particular AV comprise an AV software version of the particular AV, and wherein the likelihood of the failure by the particular AV to perform the one or more AV maneuvers at the one or more portions of the route is at least partly based on the AV software version of the particular AV.
13. The system of claim 8 , wherein the reroute likelihood is based on at least one of an AV maneuver type and traffic conditions.
14. The system of claim 8 , wherein the instructions, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
updating one or more map priors following the at least one AV maneuver.
15. A non-transitory computer-readable storage medium comprising instructions stored therein, which when executed by one or more processors, cause the one or more processors to perform operations comprising:
identifying a route between a first map location and a second map location;
obtaining, from each respective autonomous vehicle (AV) of a set of AVs, sensor data indicating whether the respective AV successfully completed an AV maneuver at one or more locations along the route;
based on one or more capabilities of a particular AV and the sensor data indicating whether the respective AV successfully completed the AV maneuver at one or more locations along the route, determining a likelihood of a failure by the particular AV to perform one or more AV maneuvers at one or more portions of the route;
based on the likelihood of the failure by the particular AV to perform the one or more AV maneuvers at the one or more portions of the route, determining a reroute likelihood associated with at least one AV maneuver from the one or more AV maneuvers;
calculating an estimated time of arrival (ETA) based on the reroute likelihood associated with at least one AV maneuver; and
sending the calculated ETA to a computing device associated with a user.
16. The non-transitory computer-readable storage medium of claim 15 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a map location.
17. The non-transitory computer-readable storage medium of claim 15 , wherein the reroute likelihood associated with the at least one AV maneuver corresponds with a lane identifier (ID).
18. The non-transitory computer-readable storage medium of claim 15 , wherein calculating the ETA is further based on a time of day.
19. The non-transitory computer-readable storage medium of claim 15 , wherein the reroute likelihood is based on traffic conditions.
20. The non-transitory computer-readable storage medium of claim 15 , wherein the reroute likelihood is based on an AV maneuver type.