Real-time determination of object metrics for trajectory planning

Patent metadata

Field	Value
Publication number	US-9272417-B2
Application number	US-201414333258-A
Country	US
Kind code	B2
Filing date	Jul 16, 2014
Priority date	Jul 16, 2014
Publication date	Mar 1, 2016
Grant date	Mar 1, 2016

Abstract

Example systems and methods may be used to determine a trajectory for moving an object using a robotic device. One example method includes determining a plurality of possible trajectories for moving an object with an end effector of a robotic manipulator based on a plurality of possible object measurements. The method may further include causing the robotic manipulator to pick up the object with the end effector. After causing the robotic manipulator to pick up the object with the end effector, the method may also include receiving sensor data from one or more sensors indicative of one or more measurements of the object. Based on the received sensor data, the method may additionally include selecting a trajectory for moving the object from the plurality of possible trajectories. The method may further include causing the robotic manipulator to move the object through the selected trajectory.

First claim

What is claimed is:

1. A method, comprising:
determining by one or more processors a plurality of predicted ranges of measurements of an identified object based on sensor data received from one or more sensors;
determining a plurality of distinct trajectories for moving the identified object from an initial location to a drop-off location with an end effector of a robotic manipulator, wherein the plurality of distinct trajectories correspond to the plurality of predicted ranges of measurements;
causing the robotic manipulator to pick up the identified object with the end effector;
after causing the robotic manipulator to pick up the identified object with the end effector, receiving additional sensor data from the one or more sensors indicative of one or more measurements of the identified object;
based on the received additional sensor data, selecting a trajectory for moving the identified object to the drop-off location from the plurality of determined distinct trajectories; and
causing the robotic manipulator to move the identified object through the selected trajectory.

2. The method of claim 1 , wherein the one or more sensors comprise at least one depth sensor mounted on the robotic manipulator and at least one depth sensor mounted at a fixed position within an environment of the identified object.

3. The method of claim 1 , wherein receiving additional sensor data from the one or more sensors indicative of one or more measurements of the identified object comprises:
causing the robotic manipulator to move one or more dimensions of the identified object in front of one or more depth sensors; and
receiving sensor data from the one or more depth sensors indicative of a depth of the identified object in the one or more dimensions.

4. The method of claim 1 , wherein the one or more sensors comprise a triangulation depth sensor mounted on the robotic manipulator, wherein the triangulation depth sensor comprises a laser beam and an offset camera to detect a depth of the identified object in one or more dimensions.

5. The method of claim 1 , wherein the end effector of the robotic manipulator comprises a gripper, and wherein the one or more sensors comprise a force-torque sensor positioned above the gripper on the robotic manipulator.

6. The method of claim 5 , further comprising receiving sensor data from the force-torque sensor indicative of at least one of a mass of the identified object, a center of gravity of the identified object, and an inertia matrix of the identified object.

7. The method of claim 5 , further comprising:
after causing the robotic manipulator to pick up the identified object with the end effector, causing the robotic manipulator to pivot the object around an edge of the object in contact with an adjacent surface;
while causing the robotic manipulator to pivot the object, using sensor data from the force-torque sensor to cause the robotic manipulator to follow a pivot trajectory that maintains contact between the edge of the object and the adjacent surface; and
based on the pivot trajectory followed by the robotic manipulator, determining a depth of the object from the edge of the object.

8. The method of claim 1 , wherein the plurality of distinct trajectories comprise a conservative trajectory portion and a plurality of aggressive trajectory portions, wherein the conservative trajectory portion is used for each of the plurality of predicted ranges of measurements and the plurality of aggressive trajectory portions correspond to the predicted ranges of measurements, the method further comprising:
determining an estimated point in time by which the additional sensor data will have been received, wherein the robotic manipulator moves the identified object from the initial location through the conservative trajectory portion prior to the estimated point in time;
determining an estimated point in space where the robotic manipulator will be at the estimated point in time as the robotic manipulator moves the identified object from the initial location through the conservative trajectory portion; and
determining the plurality of aggressive trajectory portions to begin at the estimated point in space, wherein the robotic manipulator moves the identified object from the estimated point in space to the drop-off location through the aggressive trajectory portion of the selected trajectory after the estimated point in time.

9. The method of claim 1 , wherein the plurality of distinct trajectories comprise a plurality of distinct paths of the end effector of the robotic manipulator from the initial location to the drop-off location.

10. The method of claim 1 , wherein the plurality of distinct trajectories comprise a plurality of velocity profiles of the end effector of the robotic manipulator for traversing each of the plurality of distinct trajectories.

11. The method of claim 1 , wherein the plurality of predicted ranges of measurements correspond to different predicted dimensions of the identified object.

12. A system, comprising:
a robotic manipulator;
at least one sensor; and
a control system configured to:
determine a plurality of predicted ranges of measurements of identified object based on sensor data received from the at least one sensor;
determine a plurality of distinct trajectories for moving the identified object from an initial location to a drop-off location with an end effector of the robotic manipulator, wherein the plurality of distinct trajectories corresponds to the plurality of predicted ranges of measurements;
cause the robotic manipulator to pick up the identified object with the end effector;
after causing the robotic manipulator to pick up the identified object with the end effector, receive additional sensor data from the at least one sensor indicative of one or more measurements of the identified object;
based on the received additional sensor data, select a trajectory for moving the identified object to the drop-off location from the plurality of determined distinct trajectories; and
cause the robotic manipulator to move the identified object through the selected trajectory.

13. The system of claim 12 , wherein the at least one sensor comprises at least one depth sensor mounted on the robotic manipulator and at least one depth sensor mounted at a fixed position within an environment of the identified object.

14. The system of claim 12 , wherein the at least one sensor comprises a triangulation depth sensor mounted on the robotic manipulator, wherein the triangulation depth sensor comprises a laser beam and an offset camera to detect a depth of the identified object in one or more dimensions.

15. The system of claim 12 , wherein the end effector of the robotic manipulator comprises a gripper, and wherein the at least one sensor comprises a force-torque sensor positioned above the gripper on the robotic manipulator to determine at least one of a mass of the identified object, a center of gravity of the identified object, and an inertia matrix of the identified object.

16. A non-transitory computer readable medium having stored therein instructions, that when executed by a computing system, cause the computing system to perform functions comprising:
determining a plurality of predicted ranges of measurements of an identified object based on sensor data received from one or more sensors;
determining a plurality of distinct trajectories for moving the identified object from an initial location to a drop-off location with an end effector of a robotic manipulator, wherein the plurality of distinct trajectories correspond to the plurality of

Assignees

• Google Inc

Inventors

• Konolige Kurt
• Rublee Ethan
• Kalakrishnan Mrinal

Classifications

• B25J9/1664 Primary

  characterised by motion, path, trajectory planning · CPC title

• Y10S901/30

  End effector · CPC title

• Y10S901/27

  Arm part · CPC title

• Y10S901/46

  Sensing device · CPC title

• B25J9/1612 Primary

  characterised by the hand, wrist, grip control · CPC title