Discrete decision architecture for motion planning system of an autonomous vehicle
The present disclosure provides autonomous vehicle systems and methods that include or otherwise leverage a motion planning system that generates constraints as part of determining a motion plan for an autonomous vehicle (AV). In particular, a scenario generator within a motion planning system can generate constraints based on where objects of interest are predicted to be relative to an autonomous vehicle. A constraint solver can identify navigation decisions for each of the constraints that provide a consistent solution across all constraints. The solution provided by the constraint solver can be in the form of a trajectory path determined relative to constraint areas for all objects of interest. The trajectory path represents a set of navigation decisions such that a navigation decision relative to one constraint doesn't sacrifice an ability to satisfy a different navigation decision relative to one or more other constraints.
1. A computing system for controlling an autonomous vehicle, the computing system comprising:
one or more processors; and
one or more non-transitory computer-readable media that store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
determining a plurality of multi-dimensional spaces corresponding to a maneuver associated with a nominal path of the autonomous vehicle, the plurality of multi-dimensional spaces defining spaces and corresponding times in which the autonomous vehicle operates during the maneuver;
generating, for an object of interest, a constraint area in a respective multi-dimensional space of the plurality of multi-dimensional spaces, the constraint area representing a subset of the respective multi-dimensional space in which the autonomous vehicle cannot be located at particular times;
determining a multiplexed space using the plurality of multi-dimensional spaces;
determining a path across the multiplexed space that does not interfere with the constraint area; and
controlling motion of the autonomous vehicle based at least in part on the path.
2. The computing system of claim 1 , wherein determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area comprises:
determining an applicable portion of the respective multi-dimensional space that corresponds to operation of the autonomous vehicle for a respective phase of a plurality of phases of the maneuver; and
determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area in the applicable portion.
3. The computing system of claim 2 , wherein the maneuver is a lane change maneuver associated with at least a first lane and a second lane associated with the nominal path.
4. The computing system of claim 3 , wherein:
a first phase of the plurality of phases corresponds to the first lane; and
a second phase of the plurality of phases corresponds to the second lane.
5. The computing system of claim 2 , wherein:
the respective phase is a first phase, the respective multi-dimensional space is a first multi-dimensional space, and the applicable portion is a first applicable portion; and
the operations further comprise:
determining a second applicable portion of a second multi-dimensional space that corresponds to operation of the autonomous vehicle in a second different phase; and
determining the multiplexed space comprising the first applicable portion and the second applicable portion.
6. The computing system of claim 1 , wherein the respective multi-dimensional space comprises a first dimension indicating a distance of travel along the nominal path, wherein the nominal path is defined based on map.
7. The computing system of claim 1 , wherein controlling motion of the autonomous vehicle based at least in part on the path comprises:
determining a set of navigation decisions based at least in part on the path;
determining a motion plan based at least in part the set of navigation decisions, the motion plan comprising a planned trajectory for the autonomous vehicle, the planned trajectory comprising vehicle states and controls to achieve the vehicle states; and
controlling motion of the autonomous vehicle based at least in part on executing the planned trajectory of the motion plan.
8. The computing system of claim 6 , wherein the multi-dimensional space is two-dimensional, wherein a second dimension is time.
9. A method for controlling an autonomous vehicle, the method comprising:
determining a plurality of multi-dimensional spaces corresponding to a maneuver associated with a nominal path of the autonomous vehicle, the plurality of multi-dimensional spaces defining spaces and corresponding times in which the autonomous vehicle operates during the maneuver;
generating, for an object of interest, a constraint area in a respective multi-dimensional space of the plurality of multi-dimensional spaces, the constraint area representing a subset of the respective multi-dimensional space in which the autonomous vehicle cannot be located at particular times;
determining a multiplexed space using the plurality of multi-dimensional spaces;
determining a path across the multiplexed space that does not interfere with the constraint area; and
controlling motion of the autonomous vehicle based at least in part on the path.
10. The method of claim 9 , wherein determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area comprises:
determining an applicable portion of the respective multi-dimensional space that corresponds to operation of the autonomous vehicle for a respective phase of a plurality of phases of the maneuver; and
determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area in the applicable portion.
11. The method of claim 10 , wherein the maneuver is a lane change maneuver associated with at least a first lane and a second lane associated with the nominal path.
12. The method of claim 11 , wherein:
a first phase of the plurality of phases corresponds to the first lane; and
a second phase of the plurality of phases corresponds to the second lane.
13. The method of claim 10 , wherein:
the respective phase is a first phase, the respective multi-dimensional space is a first multi-dimensional space, and the applicable portion is a first applicable portion; and
the method further comprises:
determining a second applicable portion of a second multi-dimensional space that corresponds to operation of the autonomous vehicle in a second different phase; and
determining the multiplexed space comprising the first applicable portion and the second applicable portion.
14. The method of claim 9 , wherein the respective multi-dimensional space comprises a dimension indicating a distance of travel along the nominal path, wherein the nominal path is defined based on map data.
15. The method of claim 9 , wherein controlling motion of the autonomous vehicle based at least in part on the path comprises:
determining a set of navigation decisions based at least in part on the path;
determining a motion plan based at least in part the set of navigation decisions, the motion plan comprising a planned trajectory for the autonomous vehicle, the planned trajectory comprising vehicle states and controls to achieve the vehicle states; and
controlling motion of the autonomous vehicle based at least in part on executing the planned trajectory of the motion plan.
16. The method of claim 14 , wherein the multi-dimensional space is two-dimensional, wherein a second dimension is time.
17. An autonomous vehicle, comprising:
one or more processors; and
one or more non-transitory computer-readable media that store instructions that, when executed by the one or more processors, cause the autonomous vehicle to perform operations, the operations comprising:
determining a plurality of multi-dimensional spaces corresponding to a maneuver associated with a nominal path of the autonomous vehicle, the plurality of multi-dimensional spaces defining spaces and corresponding times in which the autonomous vehicle operates during the maneuver;
generating, for an object of interest, a constraint area in a respective multi-dimensional space of the plurality of multi-dimensional spaces, the constraint area representing a subset of the respective multi-dimensional space in which the autonomous vehicle cannot be located at particular times;
determining a multiplexed space using the plurality of multi-dimensional spaces;
determining a path across the multiplexed space that does not interfere with the constraint area; and
controlling motion of the autonomous vehicle based at least in part on the path.
18. The autonomous vehicle of claim 17 , wherein determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area comprises:
determining an applicable portion of the respective multi-dimensional space that corresponds to operation of the autonomous vehicle for a respective phase of a plurality of phases of the maneuver; and
determining the path across the plurality of multi-dimensional spaces that does not interfere with the constraint area in the applicable portion.
19. The autonomous vehicle of claim 18 , wherein:
the maneuver is a lane change maneuver associated with at least a first lane and a second lane associated with the nominal path;
a first phase of the plurality of phases corresponds to the first lane; and
a second phase of the plurality of phases corresponds to the second lane.
20. The autonomous vehicle of claim 17 , wherein controlling motion of the autonomous vehicle based at least in part on the path comprises:
determining a set of navigation decisions based at least in part on the path;
determining a motion plan based at least in part the set of navigation decisions, the motion plan comprising a planned trajectory for the autonomous vehicle, the planned trajectory comprising vehicle states and controls to achieve the vehicle states; and
controlling motion of the autonomous vehicle based at least in part on executing the planned trajectory of the motion plan.