APPARATUS FOR INPUTTING TEACHING DATA AND APPARATUS AND METHOD FOR GENERATING TEACHING COMMAND FOR ROBOT

• US 20150094853A1
• Filed: 02/28/2014
• Published: 04/02/2015
• Est. Priority Date: 10/01/2013
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: LEE, Hoo Man, Kim, Joong Bae
• Granted Patent: US 9,393,690 B2
• Time in Patent Office: Days
• US Class Current: 700/250
• CPC Class Codes: B25J 9/1602 characterised by the control system, structure, architecture

Abstract

An apparatus for inputting teaching data for a robot includes a first input interface configured to input teaching data of a path layer; and a second input interface configured to input teaching data of a task layer.

Claims

1. An apparatus for inputting teaching data for a robot, the apparatus comprising:

a first input interface configured to input teaching data of a path layer; and

a second input interface configured to input teaching data of a task layer.

2. The apparatus of claim 1, wherein the first input interface comprises an exoskeleton-type hardware-based input interface.

3. The apparatus of claim 1, wherein the first input comprises:

one or more sensors configured to measure rotation angles of joints of the user;

a calculation unit configured to calculate rotation angular velocity and rotation angular acceleration of the joints using the joint rotation angles;

a first teaching data producing unit configured to produce the teaching data of the path layer based on the rotation angle, rotation angular velocity and rotation angular acceleration of the joints.

4. The apparatus of claim 3, wherein the calculation unit is configured to calculate the rotation angular velocity and rotation angular acceleration of the joints using the rotation angles of the joints that are measured by the sensors.

5. The apparatus of claim 1, wherein the second input interface comprises a vision sensor-based input interface.

6. The apparatus of claim 1, wherein the second input interface comprises:

a photographing unit configured to store an image that photographs a motion of the user;

a motion recognition unit configured to recognize the motion in the stored image; and

a second teaching data producing unit configured to produce teaching data of the task layer based on the recognized motion.

7. A control apparatus of a robot, comprising:

an input unit configured to receive teaching data of a path layer and a task layer;

a teaching data interpretation module configured to interpret the teaching data of the path layer or the task layer; and

a teaching data production module configured to produce a teaching command for controlling the robot using the interpreted teaching data by the teaching data interpretation module.

8. The control apparatus of claim 7, wherein the teaching data interpretation module comprises:

a first interpreter configured to interpret the teaching data of the task layer into a sequence of the task layer using a preset learning algorithm; and

a second interpreter configured to interpret the teaching data of the path layer into a robot path in a joint space or workspace of the task layer using a preset algorithm.

9. The control apparatus of claim 8, wherein the first interpreter is configured to reduce the dimension of the teaching data of the task layer to perform a data encoding of the teaching data and extract semantics of the task layer from the encoded teaching data of the task layer, and

wherein the teaching command producing module comprises:

a conversion unit configured to produce a semantic sequence using the semantics of the task layer; and

a task sequence producing unit configured to extract a behavior semantic suitable for the semantic sequence from a behavior database having pre-stored behaviors and produce a task sequence for the robot using the extracted behavior sequence.

10. The control apparatus of claim 8, wherein the second interpreter is configured to reduce the dimension of the teaching data of the path layer to encode the teaching data and decode the encoded teaching data to produce path information, and

wherein the teaching command producing module is configured to produce joint space path information and workspace path information by applying a preset kinematic transformation algorithm to the decoded information.

11. A method for producing a teaching command for a robot, the method comprising:

receiving teaching data of a path layer and a task layer;

interpreting the teaching data of the task layer to produce semantics of the task layer;

producing a semantic sequence using the semantics of the task layer;

extracting a behavior sequence suitable for the semantic sequence from a behavior database having pre-stored behaviors and producing a task sequence of the robot using the extracted behavior sequence; and

interpreting the teaching data of the path layer to produce a joint space path or workspace path of the path layer.

12. The method of claim 11, wherein said receiving teaching data comprises:

receiving the teaching data of the task layer through an exoskeleton-type hardware-based input interface; and

receiving the teaching data of the path layer through a vision sensor-based input interface.

13. The method of claim 12, wherein said receiving teaching data comprises:

receiving rotation angles of the joints of the user through an exoskeleton-type hardware-based input interface having at least one or more sensors;

calculating rotation angular velocities and rotation angular accelerations of the joints using the rotation angles of the joints; and

receiving the teaching data of the path layer that is produced based on the joint rotation angle, the rotation angular velocity or rotation angular acceleration.

14. The method of claim 13, wherein said receiving teaching data comprises:

recognizing a motion in an image that photographs the motion of the user; and

receiving the teaching data of the task layer that is produced based on the recognized motion.