US 12,447,614 B2
Learning from demonstration for determining robot perception motion
Ammar Husain, San Francisco, CA (US); and Mikael Persson, Sunnyvale, CA (US)
Assigned to GDM Holding LLC, Mountain View, CA (US)
Filed by GDM Holding LLC, Mountain View, CA (US)
Filed on Sep. 14, 2022, as Appl. No. 17/932,271.
Claims priority of provisional application 63/244,902, filed on Sep. 16, 2021.
Prior Publication US 2023/0084774 A1, Mar. 16, 2023
Int. Cl. B25J 19/02 (2006.01); B25J 9/16 (2006.01)
CPC B25J 9/1664 (2013.01) [B25J 9/161 (2013.01); B25J 9/163 (2013.01); B25J 9/1689 (2013.01); B25J 19/021 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
determining, for a robotic device that comprises a perception system, a robot planner state representing at least one future path for the robotic device in an environment, wherein the perception system is independently controllable to move relative to the robotic device to view the environment from different perspectives;
determining a perception system trajectory by inputting at least the robot planner state into a machine learning model trained based on training data comprising at least a plurality of robot planner states and a corresponding plurality of operator-directed perception system trajectories, wherein each of the operator-directed perception system trajectories is based on operator-generated instructions for control of a respective perception system relative to a respective robotic device for the corresponding robot planner state, and wherein the perception system trajectory comprises a trajectory through which to move the perception system relative to the robotic device; and
controlling, by the robotic device, the perception system to move through the determined perception system trajectory relative to the robotic device.