Performance recreation system
AssigneeAivot, LLC
InventorShashwat Srivastav
Life of this patent
Estimating the term…
Abstract
The present disclosure generally relates to performance recreation, and in particular, the recreation of observed human performance using reinforcement learning. In this regard, a first object is identified from a plurality of objects. The manipulation of the first object is tracked from a first position to a second position. A characterization of the manipulation is generated. A policy that controls a mechanical gripper to recreate the manipulation is generated based on an iteratively increasing cumulative award. The mechanical gripper iteratively recreates the manipulation to increase a cumulative award with each recreation.
Claims20
A robotic system, comprising:
- one or more memory devices coupled with one or more processors configured to:
- employ an arm assembly to initiate moving an object away from an initial position and towards a final position within an environment of the robotic system;
- receive, from a camera, image data depicting the arm assembly moving the object from the initial position towards the final position;
- operate the arm assembly based on the image data to move the arm assembly to the final position;
- determine an award based on a plurality of weights associated with a plurality of points, the plurality of points including the final position and at least one intermediary position between the initial position and the final position;
- and
- generate or update a policy to perform a future movement of the arm assembly through the environment based on the award determined with the plurality of weights.
A method, comprising:
- employing, by one or more processors, an arm assembly of a robotic system to initiate moving an object away from an initial position and towards a final position within an environment;
- receiving, by the one or more processors, from a camera, image data depicting the arm assembly moving the object from the initial position towards the final position;
- operating, by the one or more processors, the arm assembly based on the image data to move the arm assembly to the final position;
- determining, by the one or more processors, an award based on a plurality of weights associated with a plurality of points, the plurality of points including the final position and at least one intermediary position between the initial position and the final position;
- and
- generating or updating, by the one or more processors, a policy to perform a future movement of the arm assembly through the environment based on the award determined with the plurality of weights.
One or more non-transitory storage medium storing instructions thereon, that, when executed by one or more processors, cause the one or more processors to perform operations, the operations comprising:
- employing an arm assembly of a robotic system to initiate moving an object away from an initial position and towards a final position within an environment;
- receiving from a camera, image data depicting the arm assembly moving the object from the initial position towards the final position;
- operating the arm assembly based on the image data to move the arm assembly to the final position;
- determining an award based on a plurality of weights associated with a plurality of points, the plurality of points including the final position and at least one intermediary position between the initial position and the final position;
- and
- generating or updating a policy to perform a future movement of the arm assembly through the environment based on the award determined with the plurality of weights.
Dependencies are inferred from the claim text.