Self-calibrating orienting system for a manipulating device

• US 20050234566A1
• Filed: 04/08/2005
• Published: 10/20/2005
• Est. Priority Date: 04/16/2004
• Status: Active Grant

• Current Assignee: Honda Research Institute Europe GmbH (Honda Motor Company)
• Original Assignee: Honda Research Institute Europe GmbH (Honda Motor Company)
• Inventors: Joublin, Frank, Rodemann, Tobias
• Granted Patent: US 7,509,177 B2
• Time in Patent Office: Days
• US Class Current: 700/61
• CPC Class Codes: B25J 9/1697 Vision controlled systems
• CPC Class Codes: G05B 19/401 characterised by control arrangements for measuring, e.g. calibration and initialisation, measuring workpiece for machining purposes G05B19/19 takes precedence
• CPC Class Codes: G05B 2219/39367 Using a motion map, association between visual position and joint position
• CPC Class Codes: G05B 2219/39391 Visual servoing, track end effector with camera image feedback
• CPC Class Codes: G05B 2219/40617 Agile eye, control position of camera, active vision, pan-tilt camera, follow object

Abstract

The invention relates to the control of an orienting/positioning system comprising comprising at least a sensor and an actuator for controlling an orienting and/or positioning action adapted to change an input space of the sensor. A first step evaluates pre-action output information of the sensor in order to detect a pre-action position of a manipulating device in the input space of the sensor. A second step decides on a targeted post-action position of the manipulating device in the input space of the sensor. A third step defines a command for the actuator by mapping any deviation of the pre-action position and the targeted post-action position in the input space of the sensor to actuator control coordinates using a predefined mapping function. A fourth step orients/positions the manipulating device according to the defined command in order to carry out the orienting/positioning action. A fifth step detects a real post-action position of the manipulating device in the input space of the sensor. A sixth step adapts the mapping function used in step three based on any difference of the real post-action position and the targeted post-action position of the manipulating device in the input space of the sensor in order to carry out an incremental adaptive learning of the mapping function. According to one embodiment of the present invention, steps S 1 to S6 are cyclically repeated at least once using the respectively adapted mapping function.

Claims

1. A method for controlling an orienting/positioning system, the orienting/positioning system comprising at least a sensor and an actuator for controlling at least one of an orienting and a positioning action adapted to change an input space of the sensor, the method comprising the steps of:

(S1) evaluating pre-action output information of the sensor in order to detect a pre-action position of a manipulating device in the input space of the sensor;

(S2) determining a targeted post-action position of the manipulating device in the input space of the sensor;

(S3) defining a command for the actuator by mapping any deviation of the pre-action position and the targeted post-action position in the input space of the sensor to actuator control coordinates using a predefined mapping function;

(S4) orienting the manipulating device by the actuator according to the defined command in order to carry out the orienting action;

(S5) detecting a real post-action position of the manipulating device in the input space of the sensor; and

(S6) adapting the mapping function used in step S3 based on any difference of the real post-action position and the targeted post-action position of the manipulating device in the input space of the sensor in order to carry out an incremental adaptive learning of the mapping function, wherein steps S1 to S6 are cyclically repeated at least once using the respectively adapted mapping function.

2. The method of claim 1,wherein step S6 comprises the steps of:

calculating a correction signal based on any difference of the real post-action position and the targeted position of the manipulating device in the input space of the sensor, and

adapting the mapping function taking into account the correction signal, the executed actuator command and the pre-action position of the manipulating device in the input space of the sensor.

3. The method of claim 2, wherein before the adaptation of the mapping function, an adaptation vector is multiplied with a factor larger than or equal to 0 and smaller than or equal to 1 in order to control an adaptation rate for the mapping function.

4. The method of claim 2, further comprising the steps of:

computing a confidence value based on a plausibility of the difference of the real post-action position and the targeted position of the manipulating device; and

modulating an adaptation vector with the confidence value such that an adaptation rate depends on the confidence value.

5. The method of claim 1, wherein the mapping function is implemented by means of a look-up table.

6. The method of claim 1, wherein in steps S1 to S6, for each cycle, not only the position of the manipulating device, but also a neighborhood in a defined range surrounding the manipulating device is used.

7. The method of claim 6, wherein the range of the neighborhood can be changed between two subsequent cycles.

8. The method of claim 1, wherein the adaptation of the mapping function is carried out in the sensor coordinates.

9. The method of claim 1, wherein the adaptation of the mapping function is carried out in the actuator coordinates.

10. A computer software product adapted for implementing the method of claim 1 when running on a computing device.

11. A method for controlling an orienting/positioning system, the orienting/positioning system comprising at least a sensor and an actuator for controlling at least one of an orienting and a positioning action of a manipulating device able to change an input space of the sensor, the method comprising the steps of:

(S1) mapping output information of the sensor in the input space of the sensor to actuator control coordinates using a predefined mapping function;

(S2) determining a targeted post-action position, the decision taking place in the actuator control coordinates;

(S3) defining a command for the actuator based on a result of the decision; and

(S4) orienting the manipulating device by the actuator according to the defined command.

12. The method of claim 11, wherein outputs of a plurality of sensors are individually mapped into actuator coordinates and then combined, wherein the decision in step S2 is carried out on the basis of a result of the combination.

13. The method of claim 11, wherein constraints of the actuator are taken into account when making the decision in step S2.

14. An auto-calibrating orienting/positioning system, comprising at least a sensor means, a computing means and an actuator means for controlling at least one of an orienting and a positioning action of a manipulating device adapted to change an input space of the sensor means, the computing means comprising:

means for evaluating pre-action output information of the sensor means in order to detect a pre-action position of the manipulating device in the input space of the sensor means;

means for deciding on a targeted post-action position of the manipulating device in the input space of the sensor means;

means for commanding the actuator means, a command being generated by mapping any deviation of the pre-action position and the targeted post-action position in the input space of the sensor means to actuator control coordinates using a predefined mapping function;

means for detecting a real post-action position of the manipulating device in the input space of the sensor means; and

means for calibrating the mapping function based on any difference of the real post-action position and the targeted post-action position of the manipulating device in the input space of the sensor means in order to carry out an incremental adaptive learning of the mapping function.

15. The system of claim 14, further comprising means for calculating a correction signal based on any difference of the real post-action position and the targeted position of the manipulating device in the input space of the sensor means, wherein the mapping function is calibrated taking into account the correction signal and the pre-action position of the manipulating device in the input space of the sensor means.

16. The system of claim 15, wherein before the calibration of the mapping function, the correction signal is multiplied with a factor larger than or equal to 0 and smaller than or equal to 1 in order to control an adaptation rate for the mapping function.

17. The system of claim 14, further comprising:

means for computing a confidence value based on a plausibility of the difference of the real post-action position and the targeted position of the manipulating device, and

means for modulating an adaptation vector with the confidence value such that an adaptation rate depends on the confidence value.

18. The system of claim 14, wherein the mapping function is implemented by means of a look-up table.

19. The system of claim 14, wherein the sensor means comprise at least one of a camera or a microphone.

20. The system of claim 14, wherein the manipulating device comprises a robotic device.