Back to results
US 2026/0285327 A1Published application
Machine-learned model architecture for predicting future object trajectory
AssigneeZoox, Inc.
InventorsZhefan Ye, Sri Nitchith Akula
Life of this patent
Estimating the term…
Abstract
Predicting a future trajectory of an object may comprise using a machine-learned model architecture to iteratively predict a series of predicted poses (positions and orientations) of the object over time. The machine-learned model may condition these predictions on a lane orientation of nearest target pose to a predicted pose or last predicted pose of the object.
Claims21
(canceled)
A system comprising:
- one or more processors;
- and
- one or more non-transitory computer-readable media storing processor-executable instructions that, when executed by the one or more processors, cause the system to perform operations comprising:
- determining a current position and a first predicted pose of an object in an environment in which a vehicle is operating, wherein the first predicted pose indicates one or more of a first predicted position or a first predicted orientation of the object;
- and
- determining a predicted trajectory of the object based at least in part on:
- determining, by a machine-learned model and based at least in part on the track, the first predicted pose, and one or more of the current position or a previous position, a second predicted pose of the object indicating one or more of a second predicted position or a second predicted orientation of the object;
- and
- controlling the vehicle based at least in part on the predicted trajectory of the object.
6 more elements
One or more non-transitory computer-readable media comprising processor-executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
- determining a current position and a first predicted pose of an object in an environment in which a vehicle is operating, wherein the first predicted pose indicates at least one of a first predicted position or a first predicted orientation of the object;
- determining a predicted trajectory of the object based at least in part on:
- determining, by a machine-learned model based at least in part on the current position and the first predicted pose, a second predicted pose of the object indicating at least one of a second predicted position or a second predicted orientation of the object;
- and
- controlling the vehicle based at least in part on the predicted trajectory of the object.
2 more elements
A method comprising:
- determining a current pose and a first predicted pose of an object in an environment in which a vehicle is operating, wherein the first predicted pose indicates at least one of a first predicted position or a first predicted orientation of the object;
- determining a predicted trajectory of the object based at least in part on:
- determining, by a machine-learned model based at least in part on the current position and the first predicted pose, a second predicted pose of the object indicating at least one of a second predicted position or a second predicted orientation of the object;
- and
- controlling the vehicle based at least in part on the predicted trajectory of the object.
2 more elements
Dependencies are inferred from the claim text.