Object motion prediction and autonomous vehicle control
Systems and methods for predicting object motion and controlling autonomous vehicles are provided. In one example embodiment, a computer implemented method includes obtaining state data indicative of at least a current or a past state of an object that is within a surrounding environment of an autonomous vehicle. The method includes obtaining data associated with a geographic area in which the object is located. The method includes generating a combined data set associated with the object based at least in part on a fusion of the state data and the data associated with the geographic area in which the object is located. The method includes obtaining data indicative of a machine-learned model. The method includes inputting the combined data set into the machine-learned model. The method includes receiving an output from the machine-learned model. The output can be indicative of a plurality of predicted trajectories of the object.
1. An autonomous vehicle control system comprising:
one or more processors; and
one or more computer-readable medium storing instructions executable by the one or more processors to cause the control system to perform operations, the operations comprising:
obtaining state data indicative of at least a current state or a past state of an object that is within a surrounding environment of an autonomous vehicle;
providing input data descriptive of the state data indicative of at least the current state or the past state of the object to a machine-learned model;
receiving as an output of the machine-learned model, in response to receipt of the input data, an output indicative of a plurality of predicted trajectories of the object, the plurality of predicted trajectories comprising a plurality of predicted waypoints, wherein the plurality of predicted waypoints is associated with a waypoint confidence level indicative of per-waypoint uncertainty for the plurality of predicted waypoints; and
controlling a motion of the autonomous vehicle based on the waypoint confidence level.
2. The control system of claim 1 , wherein the operations further comprise:
displaying an indicator indicative of the waypoint confidence level for the plurality of predicted waypoints.
3. The control system of claim 1 , wherein the plurality of predicted waypoints for a predicted trajectory comprises a first waypoint and a second waypoint, wherein the first waypoint comprises a first indicator indicative of a first waypoint confidence level for the first waypoint, wherein the second waypoint comprises a second indicator indicative of a second confidence level for the second waypoint.
4. The control system of claim 1 , wherein providing the input data descriptive of the state data indicative of at least the current state or the past state of the object to the machine-learned model comprises:
obtaining data associated with a geographic area in which the object is located; and
generating the input data based on the state data and the data associated with the geographic area in which the object is located, wherein the input data comprises a combined data set associated with the object based on a fusion of the state data and the data associated with the geographic area in which the object is located.
5. The control system of claim 1 , wherein controlling the motion of the autonomous vehicle based on the waypoint confidence level comprises:
determining a trajectory confidence level for the plurality of predicted trajectories based on the waypoint confidence level;
determining a selected predicted trajectory of the object from among the plurality of predicted trajectories of the object based the trajectory confidence level for the selected predicted trajectory; and
controlling the motion of the autonomous vehicle based on the selected predicted trajectory.
6. The control system of claim 1 , wherein the plurality of predicted waypoints are respectively indicative of a predicted future location for the object.
7. The control system of claim 6 , wherein the predicted future location for the object is relative to one or more features of a geographic area in which the object is located.
8. The control system of claim 6 , wherein the plurality of predicted waypoints respectively comprise an indicator indicative of the waypoint confidence level for the predicted waypoint, and wherein the indicator indicative of the waypoint confidence level for the predicted waypoint quantifies an uncertainty of the predicted future location for the object.
9. The control system of claim 1 , wherein the machine-learned model is previously trained to determine a respective waypoint confidence level for each of a plurality of respective waypoints of a respective trajectory for a respective object based on input data corresponding to the respective object.
10. An autonomous vehicle comprising:
one or more processors; and
one or more computer-readable medium storing instructions executable by the one or more processors to cause the autonomous vehicle to perform operations, the operations comprising:
obtaining state data indicative of at least a current state or a past state of an object that is within a surrounding environment of an autonomous vehicle;
providing input data descriptive of the state data indicative of at least the current state or the past state of the object to a machine-learned model;
receiving as an output of the machine-learned model, in response to receipt of the input data, an output indicative of a plurality of predicted trajectories of the object, the plurality of predicted trajectories comprising a plurality of predicted waypoints, wherein at of the plurality of predicted waypoints is associated with a waypoint confidence level indicative of per-waypoint uncertainty for the plurality of predicted waypoints; and
controlling, a motion of the autonomous vehicle based on the waypoint confidence level.
11. The autonomous vehicle of claim 10 , wherein providing the input data descriptive of the state data indicative of at least the current state or the past state of the object to the machine-learned model comprises:
obtaining data indicative of one or more nominal pathways for the object; and
generating the input data based on the state data and the data indicative of the one or more nominal pathways for the object, wherein the input data comprises a combined data set associated with the object based on the state data and the data indicative of one or more nominal pathways for the object.
12. The autonomous vehicle of claim 10 , wherein providing the input data descriptive of the state data indicative of at least the current state or the past state of the object to the machine-learned model comprises:
obtaining data associated with a geographic area in which the object is located; and
generating the input data based on the state data and the data associated with the geographic area in which the object is located, wherein the input data comprises a combined data set associated with the object based on a fusion of the state data and the data associated with the geographic area in which the object is located.
13. The autonomous vehicle of claim 10 , wherein controlling the motion of the autonomous vehicle based on the waypoint confidence level for the plurality of predicted waypoints, comprises:
determining a trajectory confidence level for the plurality of predicted trajectories based on the waypoint confidence level;
determining a selected predicted trajectory of the object from among the plurality of predicted trajectories of the object based on the trajectory confidence level for the selected predicted trajectory; and
controlling the motion of the autonomous vehicle based on the selected predicted trajectory.
14. The autonomous vehicle of claim 10 , wherein the plurality of predicted waypoints for the predicted trajectory comprises a first waypoint and a second waypoint, wherein the first waypoint comprises a first indicator indicative of a first waypoint confidence level for the first waypoint, wherein the second waypoint comprises a second indicator indicative of a second confidence level for the second waypoint.
15. A computer-implemented method, comprising:
obtaining state data indicative of at least a current state or a past state of an object that is within a surrounding environment of an autonomous vehicle;
providing input data descriptive of the state data indicative of at least the current state or the past state of the object to a machine-learned model;
receiving as an output of the machine-learned model, in response to receipt of the input data, an output indicative of a plurality of predicted trajectories of the object, the plurality of predicted trajectories comprising a plurality of predicted waypoints, wherein the plurality of predicted waypoints is associated with a waypoint confidence level indicative of per-waypoint uncertainty for the plurality of predicted waypoints; and
controlling a motion of the autonomous vehicle based on the waypoint confidence level.
16. The computer-implemented method of claim 15 , wherein the plurality of predicted waypoints for predicted trajectory comprises a first waypoint and a second waypoint, wherein the first waypoint comprises a first indicator indicative of a first waypoint confidence level for the first waypoint, wherein the second waypoint comprises a second indicator indicative of a second confidence level for the second waypoint.
17. The computer-implemented method of claim 15 , wherein the machine-learned model is previously trained to determine a respective waypoint confidence level for each of a plurality of respective waypoints of a respective trajectory for a respective object based on input data corresponding to the respective object.
18. The computer-implemented method of claim 17 , wherein the machine-learned model is previously trained based on training data, wherein the training data is associated with a plurality of previously observed trajectories for a plurality of objects.