Granted Patent B2

US 12,233,916 · App. 18/065,031 · Granted Feb 25, 2025

Inventors: Randall Schur (Pittsburgh, PA); Constantin Savtchenko (Sewickley, PA)

Assignee: Ford Global Technologies, LLC

B60W60/0027B60W60/0011

Abstract

Claims (37)

1. A method for operating a vehicle, comprising:

using, by a processor, a plurality of kinematic models to generate forecasted trajectories of an actor, the plurality of kinematic models being respectively associated with different actor types that are assigned to an actor detected in an environment of the vehicle;

obtaining, by the processor, levels of divergence for the forecasted trajectories from reference paths;

selecting, by the processor, a first kinematic model from the plurality of kinematic models based on the forecasted trajectories and a kinematic state of the actor, wherein a level of divergence associated with the forecasted trajectory generated using the first kinematic model is smallest one of the levels of divergences;

using, by the processor, the first kinematic model to predict a first path for the actor;

selecting, by the processor, a second kinematic model from the plurality of kinematic models responsive to movement of the actor no longer being consistent with typical movement of an object of one of the different actor types that is associated with the first kinematic model;

using, by the processor, the second kinematic model to predict a second path for the actor; and

controlling, by the processor, operations of the vehicle based on the first and second paths.

2. The method according to claim 1 , further comprising, prior to using the plurality of kinematic models to generate the forecasted trajectories, filtering the plurality of kinematic models to remove at least one kinematic model from the plurality of kinematic models that is associated with one of the different actor types with a lowest assigned probability for the actor.

3. The method according to claim 1 , wherein the selecting the first kinematic model further comprises obtaining each one of the reference paths based on the kinematic state of the actor and a respective one of the different actor types.

4. The method according to claim 1 , wherein the selecting the first kinematic model comprises identifying one of the forecasted trajectories which most closely matches an observed trajectory of the actor.

5. The method according to claim 1 , wherein the selecting the first kinematic model comprises creating a hybrid kinematic model that includes a path follower algorithm of one of the plurality of kinematic models and a kinematics evaluator algorithm of another different one of the plurality of kinematic models.

6. The method according to claim 5 , wherein the path following algorithm or the kinematics evaluator algorithm comprises part of the first kinematic model.

7. The method according to claim 6 , wherein the using the first kinematic model to predict the first path for the actor comprising using the hybrid kinematic model to predict the first path for the actor.

8. The method according to claim 5 , further comprising updating the path follower algorithm or the kinematics evaluator algorithm of the hybrid kinematic model based on results from evaluating predictions made by the hybrid kinematic model to actual actions of the actor.

9. A system, comprising:

a processor;

a non-transitory computer-readable storage medium comprising programming instructions that are configured to cause the processor to implement a method for operating a vehicle, wherein the programming instructions comprise instructions to:

use a plurality of kinematic models to generate forecasted trajectories of an actor, the plurality of kinematic models being respectively associated with different actor types that are assigned to an actor detected in an environment of the vehicle;

select a first kinematic model from the plurality of kinematic models based on the forecasted trajectories and a kinematic state of the actor, the first kinematic model is selected based on levels of divergence for the forecasted trajectories from reference paths and a level of divergence associated with the forecasted trajectory generated using the first kinematic model is smallest one of the levels of divergence;

use the first kinematic model to predict a first path for the actor;

select a second kinematic model from the plurality of kinematic models responsive to movement of the actor no longer being consistent with typical movement of an object of one of the different actor types that is associated with the first kinematic model;

use the second kinematic model to predict a second path for the actor; and control operations of the vehicle based on the first and second paths.

10. The system according to claim 9 , wherein the programming instructions further comprise instructions to, prior to using the plurality of kinematic models to generate the forecasted trajectories, filter the plurality of kinematic models to remove at least one kinematic model from the plurality of kinematic models that is associated with one of the different actor types with a lowest assigned probability for the actor.

11. The system according to claim 9 wherein each one of the reference paths is obtained based on the kinematic state of the actor and a respective one of the different actor types.

12. The system according to claim 9 , wherein the first kinematic model is selected by identifying one of the forecasted trajectories which most closely matches an observed trajectory of the actor.

13. The system according to claim 9 , wherein the first kinematic model is selected by creating a hybrid kinematic model that includes a path follower algorithm of one of the plurality of kinematic models and a kinematics evaluator algorithm of another different one of the plurality of kinematic models.

14. The system according to claim 13 , wherein the path following algorithm or the kinematics evaluator algorithm comprises part of the first kinematic model.

15. The system according to claim 14 , wherein the hybrid kinematic model is used to predict the first path for the actor.

16. A non-transitory computer-readable medium that stores instructions that, when executed by at least one computing device, will cause the at least one computing device to perform operations comprising:

using a plurality of kinematic models to generate forecasted trajectories of an actor, the plurality of kinematic models being respectively associated with different actor types that are assigned to an actor detected in an environment of a vehicle;

selecting a first kinematic model from the plurality of kinematic models based on the forecasted trajectories and a kinematic state of the actor and by creating a hybrid kinematic model that includes a path follower algorithm of one of the plurality of kinematic models and a kinematics evaluator algorithm of another different one of the plurality of kinematic models;

using the first kinematic model to predict a first path for the actor;

selecting a second kinematic model from the plurality of kinematic models responsive to movement of the actor no longer being consistent with typical movement of an object of one of the different actor types that is associated with the first kinematic model;

using the second kinematic model to predict a second path for the actor;

controlling operations of the vehicle based on the first and second paths; and

updating the path follower algorithm or the kinematics evaluator algorithm of the hybrid kinematic model based on results from evaluating predictions made by the hybrid kinematic model to actual actions of the actor.

Assignments (2)

Continuity (2)

References Cited (19)