US 11,752,625 B2
Robot grip detection using non-contact sensors
Jeffrey Bingham, Sunnyvale, CA (US); Taylor Alexander, Mountain View, CA (US); Bianca Homberg, Mountain View, CA (US); and Joseph DelPreto, Los Altos, CA (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by X Development LLC, Mountain View, CA (US)
Filed on Aug. 28, 2020, as Appl. No. 17/6,274.
Application 17/006,274 is a continuation of application No. 15/839,109, filed on Dec. 12, 2017, granted, now 10,792,809.
Prior Publication US 2020/0391378 A1, Dec. 17, 2020
Int. Cl. B25J 9/16 (2006.01); B25J 13/08 (2006.01); B25J 15/08 (2006.01); B25J 15/02 (2006.01); B25J 9/10 (2006.01); B25J 19/02 (2006.01)
CPC B25J 9/1612 (2013.01) [B25J 9/1065 (2013.01); B25J 9/161 (2013.01); B25J 9/1633 (2013.01); B25J 13/082 (2013.01); B25J 15/022 (2013.01); B25J 15/0213 (2013.01); B25J 15/08 (2013.01); B25J 19/023 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
controlling a robotic gripping device to cause a plurality of digits of the robotic gripping device to move towards each other in an attempt to grasp an object;
receiving, from a first non-contact sensor, first sensor data indicative of a first region between the plurality of digits of the robotic gripping device, wherein the first sensor data is based on a first sensing modality and is indicative of the first region after the attempt to grasp the object;
receiving, from a second non-contact sensor, second sensor data indicative of a second region between the plurality of digits of the robotic gripping device, wherein the second sensor data is based on a second sensing modality than different from the first sensing modality. wherein the second sensor data is indicative of the second region after the attempt to grasp the object; and
determining, using an object-in-hand classifier that takes as input the first sensor data indicative of the first region after the attempt to grasp the object and the second sensor data indicative of the second region after the attempt to grasp the object and outputs a decision of whether the robotic gripping device is currently holding the object after the attempt to grasp the object, whether the robotic gripping device is currently holding the object after the attempt to grasp the object, wherein the object-in-hand classifier has been trained based on a training data set, wherein the training data set comprises sensor data of both the first sensing modality and the second sensing modality collected after past grasp attempts and corresponding labels of whether the past grasp attempts were successful.