Object interaction prediction systems and methods for autonomous vehicles

1. A computing system, comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause the computing system to perform operations comprising:

obtaining data associated with a first object and one or more second objects within a surrounding environment of an autonomous vehicle;

determining an interaction between the first object and the one or more second objects based at least in part on the data associated with the first object and the one or more second objects, wherein the interaction between the first object and the one or more second objects is determined based at least in part on map data associated with the surrounding environment of the autonomous vehicle;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects; and

outputting data indicative of the one or more predicted trajectories of the first object.

2. The computing system of claim 1 , wherein the data associated with the first object and the one or more second objects within the surrounding environment comprises data indicative of an initial predicted trajectory of the first object within the surrounding environment.

3. The computing system of claim 2 , further comprising:

obtaining state data indicative of one or more current or past states of the first object within the surrounding environment; and

determining the initial predicted trajectory of the first object based at least in part on the state data indicative of the one or more current or past states of the first object within the surrounding environment.

4. The computing system of claim 2 , wherein determining the interaction between the first object and the one or more second objects comprises:

determining the interaction between the first object and the one or more second objects based at least in part on the initial predicted trajectory of the first object.

5. The computing system of claim 1 , wherein the operations further comprise:

determining a probability for each of the respective one or more predicted trajectories, wherein the probability is indicative of a likelihood of the first object acting in accordance with the respective predicted trajectory.

6. The computing system of claim 1 , wherein determining the one or more predicted trajectories of the first object within the surrounding environment comprises:

obtaining data indicative of a machine-learned interaction prediction model;

inputting the data associated with the first object and the one or more second objects into the machine-learned interaction prediction model; and

obtaining an output from the machine-learned interaction prediction model, wherein the output is indicative of the one or more predicted trajectories of the first object within the surrounding environment.

7. The computing system of claim 1 , wherein the interaction is associated with a traffic rule.

8. The computing system of claim 1 , wherein the one or more second objects include a static object within the surrounding environment.

9. The computing system of claim 1 , wherein the one or more second objects include an actor object within the surrounding environment.

10. The computing system of claim 1 , wherein the one or more second objects include the autonomous vehicle.

11. The computing system of claim 1 ,

wherein determining the interaction between the first object and the one or more second objects comprises determining that a first predicted trajectory of the first object is in conflict with one or more second predicted trajectories of the one or more second objects, and

wherein determining the one or more predicted trajectories of the first object within the surrounding environment comprises in response to determining that the first predicted trajectory of the first object is in conflict with the one or more second predicted trajectories of the one or more second objects, determining the one or more predicted trajectories of the first object such that the one or more predicted trajectories of the first object are not in conflict with the one or more second predicted trajectories of the one or more second objects.

12. The computing system of claim 1 ,

wherein determining the interaction between the first object and the one or more second objects comprises associating the first object with the one or more second objects using a graph model, and

wherein determining one or more predicted trajectories of the first object within the surrounding environment comprises after associating the first object with the one or more second objects, determining the one or more predicted trajectories of the first object based on the graph model.

13. An autonomous vehicle comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause a computing system to perform operations comprising:

obtaining state data indicative of one or more current or past states of a first object and one or more second objects within a surrounding environment;

determining an initial predicted trajectory of the first object within the surrounding environment based at least in part on the state data indicative of the one or more current or past states of the first object;

determining an interaction between the first object and the one or more second objects based at least in part on the initial predicted trajectory of the first object, wherein the interaction between the first object and the one or more second objects is determined based at least in part on map data associated with the surrounding environment; and

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects.

14. The autonomous vehicle of claim 13 , wherein the operations further comprise:

determining a motion plan for the autonomous vehicle based at least in part on the one or more predicted trajectories of the first object within the surrounding environment; and

causing the autonomous vehicle to initiate a motion in accordance with at least a portion of the motion plan.

15. The autonomous vehicle of claim 13 , wherein the operations further comprise:

determining a probability for each of the respective one or more predicted trajectories, wherein the probability for the respective predicted trajectory is indicative of a likelihood of the first object acting in accordance with the respective predicted trajectory.

16. The autonomous vehicle of claim 15 , wherein the operations further comprise:

determining a score for each of the one or more predicted interaction trajectories based at least in part on the probability for each of the respective one or more predicted interaction trajectories.

17. The autonomous vehicle of claim 13 , wherein determining the interaction between the first object and the one or more second objects comprises:

determining the interaction between the first object and the one or more second objects based at least in part on a machine-learned model.

18. A computer-implemented method for determining object motion, comprising:

obtaining, by a computing system comprising one or more computing devices, data indicative of an initial predicted trajectory of a first object within a surrounding environment of an autonomous vehicle;

determining, by the computing system, an interaction between the first object and one or more second objects based at least in part on the initial predicted trajectory of the first object within the surrounding environment of the autonomous vehicle, wherein the interaction between the first object and the one or more second objects is determined based at least in part on map data associated with the surrounding environment of the autonomous vehicle;

determining, by the computing system, one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects; and

outputting, by the computing system, data indicative of the one or more predicted trajectories of the first object.

19. The computer-implemented method of claim 18 , wherein determining, by the computing system, the one or more predicted trajectories of the first object comprises:

iteratively determining, by the computing system, the one or more predicted trajectories of the first object within the surrounding environment; and

determining for each of the one or more predicted trajectories, by the computing system, a likelihood that the first object will act in accordance with the respective predicted trajectory.

20. A computing system, comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause the computing system to perform operations comprising:

obtaining data associated with a first object and one or more second objects within a surrounding environment of an autonomous vehicle;

determining an interaction between the first object and the one or more second objects based at least in part on the data associated with the first object and the one or more second objects;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects;

determining a probability for each of the respective one or more predicted trajectories, wherein the probability is indicative of a likelihood of the first object acting in accordance with the respective predicted trajectory; and

outputting data indicative of the one or more predicted trajectories of the first object.

21. A computing system, comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause the computing system to perform operations comprising:

obtaining data associated with a first object and one or more second objects within a surrounding environment of an autonomous vehicle;

determining an interaction between the first object and the one or more second objects based at least in part on the data associated with the first object and the one or more second objects;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects wherein determining the one or more predicted trajectories of the first object within the surrounding environment comprises:

obtaining data indicative of a machine-learned interaction prediction model,

inputting the data associated with the first object and the one or more second objects into the machine-learned interaction prediction model, and

obtaining an output from the machine-learned interaction prediction model, wherein the output is indicative of the one or more predicted trajectories of the first object within the surrounding environment; and

outputting data indicative of the one or more predicted trajectories of the first object.

22. A computing system, comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause the computing system to perform operations comprising:

obtaining data associated with a first object and one or more second objects within a surrounding environment of an autonomous vehicle;

determining an interaction between the first object and the one or more second objects based at least in part on the data associated with the first object and the one or more second objects, wherein determining the interaction between the first object and the one or more second objects comprises determining that a first predicted trajectory of the first object is in conflict with one or more second predicted trajectories of the one or more second objects;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects, wherein determining the one or more predicted trajectories of the first object within the surrounding environment comprises in response to determining that the first predicted trajectory of the first object is in conflict with the one or more second predicted trajectories of the one or more second objects, determining the one or more predicted trajectories of the first object such that the one or more predicted trajectories of the first object are not in conflict with the one or more second predicted trajectories of the one or more second objects; and

outputting data indicative of the one or more predicted trajectories of the first object.

23. A computing system, comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause the computing system to perform operations comprising:

obtaining data associated with a first object and one or more second objects within a surrounding environment of an autonomous vehicle;

determining an interaction between the first object and the one or more second objects based at least in part on the data associated with the first object and the one or more second objects, wherein determining the interaction between the first object and the one or more second objects comprises associating the first object with the one or more second objects using a graph model;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects, wherein determining one or more predicted trajectories of the first object within the surrounding environment comprises after associating the first object with the one or more second objects, determining the one or more predicted trajectories of the first object based on the graph model; and

outputting data indicative of the one or more predicted trajectories of the first object.

24. An autonomous vehicle comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause a computing system to perform operations comprising:

obtaining state data indicative of one or more current or past states of a first object and one or more second objects within a surrounding environment;

determining an initial predicted trajectory of the first object within the surrounding environment based at least in part on the state data indicative of the one or more current or past states of the first object;

determining an interaction between the first object and the one or more second objects based at least in part on the initial predicted trajectory of the first object;

determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects; and

determining a probability for each of the respective one or more predicted trajectories, wherein the probability for the respective predicted trajectory is indicative of a likelihood of the first object acting in accordance with the respective predicted trajectory.

25. An autonomous vehicle comprising:

one or more processors; and

one or more tangible, non-transitory, computer readable media that collectively store instructions that when executed by the one or more processors cause a computing system to perform operations comprising:

obtaining state data indicative of one or more current or past states of a first object and one or more second objects within a surrounding environment;

determining an initial predicted trajectory of the first object within the surrounding environment based at least in part on the state data indicative of the one or more current or past states of the first object;

determining an interaction between the first object and the one or more second objects based at least in part on the initial predicted trajectory of the first object; and

determining a score for each of the one or more predicted interaction trajectories based at leak in part on the probability for each of the respective one or more predicted interaction trajectories.

26. A computer-implemented method for determining object motion, comprising:

obtaining, by a computing system comprising one or more computing devices, data indicative of an initial predicted trajectory of a first object within a surrounding environment of an autonomous vehicle;

determining, by the computing system, an interaction between the first object and one or more second objects based at least in part on the initial predicted trajectory of the first object within the surrounding environment of the autonomous vehicle;

determining, by the computing system, one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects, wherein determining the one or more predicted trajectories of the first object comprises:

iteratively determining, by the computing system, the one or more predicted trajectories of the first object within the surrounding environment, and

determining for each of the one or more predicted trajectories, by the computing system, a likelihood that the first object will act in accordance with the respective predicted trajectory; and

outputting, by the computing system, data indicative of the one or more predicted trajectories of the first object.