System and method for generating a planned path using a phantom vehicle
Methods and systems that use a phantom vehicle to help generate a planned path for a real-world vehicle are described. The system will identify a starting point and a destination for a trip of the real-world vehicle. The system will select, from the data store of vehicle profiles, a phantom vehicle having an associated motion planning system that corresponds to a system that is deployed on the real-world vehicle. The system will use a high definition map to generate a planned route for the real-world vehicle from the starting point to the destination in the map. The system will run a simulation in which the phantom vehicle moves along the planned route in the map. The system will then output a record of the simulation to a user of the real world-vehicle or to a system of the real-world vehicle.
1. A system for generating a planned path for a vehicle, the system comprising:
a cloud service provider processor, and programming instructions that are configured to cause the cloud service provider processor to:
receive a trip request;
generate a planned route for a real-world vehicle in response to the trip request;
simulate a trip by a virtual vehicle by:
selecting the virtual vehicle from a plurality of different virtual vehicles profiled in a data store by characteristics of real-world vehicles, the virtual vehicle that is selected has an associated motion planning system that corresponds to a motion planning system that is deployed on the real-world vehicle;
obtaining a simulation route for the virtual vehicle based on the planned route for the real-world vehicle;
causing the virtual vehicle to make the trip by moving along the simulation route in a simulation environment to obtain first simulated vehicle perception data;
modifying the simulation route based on the first simulated vehicle perception data; and
causing the virtual vehicle to finish the trip by moving along the modified simulation route in the simulation environment to obtain second simulated vehicle perception data;
change the planned route based on at least one of the first and second simulated vehicle perception data; and
output trip instructions that are configured to cause the real-world vehicle to move along the planned route which has been changed.
2. The system of claim 1 , wherein:
a user of the real-world vehicle is a first tenant who shares use of the real-world vehicle with one or more other tenants; and
the cloud service provider processor is further caused to transmit a record of simulation to the first tenant.
3. The system of claim 1 , wherein the instructions to cause the virtual vehicle to move along the simulation route in the simulation environment comprise instructions to:
access a first tenant-specific policy for the first tenant; and
cause the virtual vehicle to operate according to the first tenant-specific policy while moving in the simulation environment.
4. The system of claim 3 , further comprising additional programming instructions that are configured to cause the cloud service provider processor to:
access a second tenant-specific policy for a second tenant;
use a high definition map to generate a second planned route for the real-world vehicle;
cause the virtual vehicle to move along the second planned route in the simulation environment according to the second tenant-specific policy; and
output a record of simulation to the second tenant.
5. The system of claim 2 , further comprising additional programming instructions that are configured to cause the cloud service provider processor to, after transmitting the record of simulation to the first tenant and in response to receiving a modification request from the first tenant, using the modification request to further modify the planned route.
6. The system of claim 1 , wherein the instructions to generate the planned route for the real-world vehicle comprise instructions to:
generate a plurality of candidate trajectories for the real-world vehicle from the starting point to the destination in a map; and
select the planned route from the plurality of candidate trajectories.
7. The system of claim 1 , wherein:
the cloud service provider system comprises a data store of a plurality of candidate motion planning systems, each of which is associated with one or more vehicles or one or more fleets; and
the programming instructions are further configured to, before generating the planned route for the real-world vehicle:
identify, from the data store, a candidate motion planning system that is associated with the real-world vehicle or one of the fleets of which the real-world vehicle is a part, and
when selecting the virtual vehicle, selecting the virutal vehicle that is associated with a motion planning system that matches the identified candidate motion planning system.
8. The system of claim 1 , wherein the programming instructions to cause the virtual vehicle to move along the simulation route in the simulation environment comprise instructions to cause a phantom vehicle service to, on a periodic basis, report a current location of the virtual vehicle.
9. The system of claim 1 , wherein the programming instructions to generate the planned route for the real-world vehicle comprise instructions to:
use one or more functions of the associated motion planning system to generate a plurality of candidate trajectories for the real-world vehicle based in part on:
one or more physical capabilities of the real-world vehicle, or
one or more operating restrictions that correspond to a trip service request.
10. A method of generating a planned path for a vehicle, the method comprising:
by a processor that is remote from a real-world vehicle:
receiving a trip request;
generating a planned route for a real-world vehicle in response to the trip request;
simulating a trip by a virtual vehicle by:
selecting the virtual vehicle from a plurality of different virtual vehicles profiled in a datastore by characteristics of real-world vehicles, the virtual vehicle that is selected has an associated motion planning system that corresponds to a motion planning system that is deployed on the real-world vehicle;
obtaining a simulation route for the virtual vehicle based on the planned route for the real-world vehicle;
causing the virtual vehicle to make the trip by moving along the simulation route in a simulation environment to obtain first simulation perception data;
modifying the simulation route based on the first simulated vehicle perception data; and
causing the virtual vehicle to finish the trip by moving along the modified simulation route in the simulation environment to obtain second simulated vehicle perception data;
changing the planned route based on at least one of the first and second simulated vehicle perception data; and
outputting trip instructions that are configured to cause the real-world vehicle to move along the planned route which has been changed.
11. The method of claim 10 , wherein:
a user of the real-world vehicle is a first tenant who shares use of the real-world vehicle with one or more other tenants; and
a record of simulation is transmitted to the first tenant.
12. The method of claim 10 , further comprising:
accessing a second tenant-specific policy for a second tenant;
using a high definition map to generate a second planned route for the real-world vehicle;
causing the virtual vehicle to move along the second planned route in the simulated environment according to the second tenant-specific policy; and
outputting a record of simulation to the second tenant.
13. The method of claim 11 , further comprising, after transmitting the record of simulation to the first tenant and in response to receiving a modification request from the first tenant, using the modification request to further modify the planned route.
14. The method of claim 10 , wherein generating the planned route for the real-world vehicle comprises:
generating a plurality of candidate trajectories for the real-world vehicle from the starting point to the destination in the map; and
selecting the planned route from the plurality of candidate trajectories.
15. The method of claim 10 , further comprising:
accessing a data store of a plurality of candidate motion planning systems, each of which is associated with one or more vehicles or one or more fleets; and
before generating the planned route for the real-world vehicle:
identifying, from the data store, a candidate motion planning system that is associated with the first real-world or one of the fleets of which the real-world vehicle is a part, and
when selecting the virtual vehicle, selecting the virtual vehicle that is associated with a motion planning system that matches the identified candidate motion planning system.
16. The method of claim 10 , wherein causing the virtual vehicle to move along the simulation route in the simulation environment comprises, by a phantom vehicle service on a periodic basis, reporting a current location of the virtual vehicle to the cloud deployment service.
17. The method of claim 10 , wherein the planned route for the real-world vehicle is generated using one or more functions of the associated motion planning system to generate a plurality of candidate trajectories for the real-world vehicle based in part on one or more of the following:
one or more physical capabilities of the real-world vehicle, or
one or more operating restrictions that correspond to a trip service request.
18. A computer program product, comprising:
a memory and programming instructions that are configured to cause a cloud service provider processor that is remote from a real-world vehicle to:
obtain a planned route for the real-world vehicle generated in response to a trip request;
simulate a trip by a virtual vehicle by:
selecting the virtual vehicle from a plurality of different virtual vehicles profiled in a data store by characteristics of real-world vehicles, the virtual vehicle that is selected has an associated motion planning system that corresponds to a motion control system that is deployed on the real-world vehicle;
obtaining a simulation route for the virtual vehicle based on the planned route for the real-world vehicle; and
causing the selected virtual vehicle to move along the simulation route in a simulation environment to obtain first simulated vehicle perception data;
modifying the simulation route based on the first simulated vehicle perception data; and
causing the virtual vehicle to finish the trip by moving along the modified simulated route in the simulation environment to obtain second simulated vehicle perception data;
change the planned route based on at least one of the first and second simulated vehicle perception data; and
output trip instructions that are configured to cause the real-world vehicle to move along the planned route that has been changed.