System and method for teaching a robot program
AssigneeScalable Robotics Inc.
InventorsThomas Andrew Fuhlbrigge, Carlos Martinez, Gregory Rossano
Life of this patent
Estimating the term…
Abstract
A system and method for teaching a robot program based on user input. In some embodiments, poses of a posing device manipulated by a user are captured and analyzed. In some embodiments, robot instructions are determined by a machine-learning algorithm that is trained on data concerning previously executed instructions evaluated by a user. In some embodiments, a user-selected process type defines workpiece geometry types for guiding the creation of robot instructions. In some embodiments, an augmented- or virtual-reality user interface is provided for obtaining user input.
Claims12
A method for teaching a robot program, the method comprising:
- obtaining geometric information about a workpiece via a 2D or 3D scanner scanning the workpiece;
- obtaining at least one pose, of a posing device, for a robot comprising a jointed arm and an end effector;
- obtaining, by user input, program information describing at least one of a motion for the robot and a behavior for the end effector of the robot, and associating the at least one of the motion and the behavior with the at least one pose;
- determining at least one best fit geometry of at least a portion of the geometric information about the workpiece;
- obtaining, by user input, a type of geometric relationship;
- evaluating the type of geometric relationship between the at least one best fit geometry and the at least one pose and displaying a representation of the geometric relationship during the obtaining of the at least one pose;
- and
- displaying a closest corresponding point of the at least one pose on the best fit geometry;
- determining an original set of instructions of the robot program based on the geometric information, the program information, and the at least one pose, wherein the original set of instructions comprises at least one of a pose for the robot, a motion for the robot, and a behavior for the end effector.
Dependencies are inferred from the claim text.