Route optimization for autonomous driving systems
Aspects and implementations of the present disclosure address shortcomings of the existing technology by enabling routing of an autonomous vehicles (AV) by identifying routes from a first location to a second location, identifying a target efficiency value of autonomous driving along a respective route, determining, in view of historical data for the respective route and using one or more randomized conditions, a confidence level associated with the target efficiency value, selecting, based on the target efficiency values and the associated confidence level, a preferred route, and causing the AV to select the first route for travel from the first location to the second location.
1. A method for routing a vehicle, the method comprising:
identifying a plurality of routes;
performing, by a processing device, for each of the plurality of routes, a plurality of simulations, wherein each of the plurality of simulations comprises:
computing a respective cost value of a plurality of cost values, the respective cost value computed based on one or more route conditions probabilistically sampled from a distribution associated with historical data for a respective route;
determining, by the processing device, using the plurality of cost values, an efficiency value for each of the plurality of routes;
selecting, based on the efficiency values, a first route of the plurality of routes; and
causing, by the processing device, the vehicle to follow the first route.
2. The method of claim 1 , wherein the one or more route conditions comprise at least one of:
randomized weather conditions along a respective route, or randomized traffic conditions along the respective route.
3. The method of claim 1 , further comprising:
determining, in view of the plurality of simulations, a confidence level for each of the plurality of routes, and wherein selecting the first route of the plurality of routes is further based on the confidence levels.
4. The method of claim 1 , wherein the one or more simulations performed for a respective route comprise characteristics of the respective route comprising at least one of:
a plurality of grade values for the respective route,
a plurality of curvature values for the respective route, speed limit information, or lane information for the respective route.
5. The method of claim 4 , wherein the characteristics of the respective route comprise an expected time of departure for the respective route.
6. The method of claim 1 , wherein the historical data comprises one or more of:
weather data for a respective route, or
traffic data for the respective route.
7. The method of claim 1 , wherein the efficiency value is based, at least in part, on one or more of:
a fuel consumption for the respective route, or
a cost of consumable components for the respective route.
8. The method of claim 1 , further comprising:
identifying an optimal trajectory for the first route, wherein the optimal trajectory comprises driving settings for the vehicle, wherein the driving settings are based on technical characteristics of the vehicle and comprise, for at least a portion of the first route, at least one of:
throttle control settings for the vehicle, or
gear control settings for the vehicle.
9. A method for routing a vehicle, the method comprising:
identifying, by a processing device, a plurality of routes;
for each route of the plurality of routes:
performing, by the processing device, a plurality of simulations to determine location-dependent vehicle settings for a plurality of locations along a respective route, wherein the location-dependent vehicle settings comprise one or more of location-dependent throttle settings, location-dependent gear settings, or location-dependent brake settings, and wherein performing a respective simulation of the plurality of simulations comprises computing, in view of the location-dependent vehicle settings, a cost value based on one or more route conditions probabilistically sampled from a distribution associated with historical data for the respective route; and
computing, using the cost values, an efficiency value for the respective route;
selecting, based on the efficiency values computed for the plurality of routes, a first route of the plurality of routes; and
causing, by the processing device, the vehicle to follow the first route.
10. The method of claim 9 , further comprising:
determining, in view of historical data for the respective route, a confidence level associated with the respective route, wherein selecting the first route is further based on the confidence levels.
11. The method of claim 9 , wherein the one or more route conditions comprise at least one of:
randomized weather conditions along the respective route, or randomized traffic conditions along the respective route.
12. The method of claim 9 , wherein the one or more simulations performed for the respective route comprise characteristics of the respective route comprising at least one of:
a plurality of grade values for the respective route,
a plurality of curvature values for the respective route, speed limit information, or
lane information for the respective route.
13. The method of claim 9 , wherein the efficiency value is based, at least in part, on one or more of:
a fuel consumption for the respective route, or
a cost of consumable components for the respective route.
14. A system for routing a vehicle, the system comprising:
a memory device; and
a processing device, communicatively coupled to the memory device, wherein the processing device is to:
identify a plurality of routes;
for each route of the plurality of routes:
perform a plurality of simulations to determine location-dependent vehicle settings for a plurality of locations along a respective route, wherein the location-dependent vehicle settings comprise one or more of location-dependent throttle settings, location-dependent gear settings, or location-dependent brake settings, and wherein a respective simulation of the plurality of simulations comprises computing, in view of the location-dependent vehicle settings, a cost value based on one or more route conditions probabilistically sampled, for the respective route, from a distribution associated with historical data; and
compute, using the cost values, an efficiency value for the respective route;
select, based on the efficiency values computed for the plurality of routes, a first route of the plurality of routes; and
cause the vehicle to follow the first route.
15. The system of claim 14 , wherein the processing device is further to:
determine, in view of historical data for the respective route, a confidence level associated with the respective route, wherein selecting the first route is further based on the confidence levels.
16. The system of claim 14 , wherein the one or more route conditions comprise at least one of:
randomized weather conditions along the respective route, or
randomized traffic conditions along the respective route.
17. The system of claim 14 , wherein the one or more simulations performed for the respective route comprise characteristics of the respective route comprising at least one of:
a plurality of grade values for the respective route,
a plurality of curvature values for the respective route, speed limit information, or
lane information for the respective route.
18. The system of claim 14 , wherein the efficiency value is based, at least in part, on one or more of:
a fuel consumption for the respective route, or
a cost of consumable components for the respective route.