Behavior control apparatus and method

• US 20040162647A1
• Filed: 01/16/2004
• Published: 08/19/2004
• Est. Priority Date: 07/16/2001
• Status: Active Grant

• Current Assignee: Honda Giken Kogyo Kabushiki Kaisha (Honda Motor Company)
• Original Assignee: Honda Giken Kogyo Kabushiki Kaisha (Honda Motor Company)
• Inventors: Tsujino, Hiroshi, Koshizen, Takamasa, Ono, Hideaki
• Granted Patent: US 7,054,724 B2
• Time in Patent Office: Days
• US Class Current: 701/2
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00

Abstract

The invention relates to a behavior control apparatus and method for autonomously controlling a mobile unit based on visual information in practical application without the needs of a greatdeal of preparation or computational cost and limiting the type of target object. According to one aspect of the invention, a method for controlling behavior of a mobile unit using behavior command is provided. First, sensory inputs are captured and then the motion of the mobile unit is estimated. The portion which includes a target object to be target for behavior of the mobile unit is segregated from the sensory inputs. The target objects extracted from the segregated portion and the location of the target object is acquired. Finally, the mobile unit is controlled based on the location of target object.

Claims

1. A behavior control apparatus for controlling behavior of a mobile unit, comprising:

sensory input capturing method for capturing sensory inputs;

motion estimating method for estimating motion of the mobile unit;

target segregation method for segregating the portion which includes a target object to be target for behavior of the mobile unit from sensory inputs;

target object matching method for extracting target object from said segregated portion;

target location acquiring method for acquiring the location of the target object;

behavior decision method for deciding behavior command for controlling the mobile unit based on the location of target object.

2. The behavior control apparatus claimed in claim 1, said motion estimating method comprising:

behavior command output method for outputting said behavior command;

behavior evaluation method for evaluating the result of the behavior of the mobile unit;

learning method for learning the motion of the mobile unit using the relationship between said sensory inputs and said behavior result; and

storage method for storing the learning result.

3. The behavior control apparatus claimed in claim 2, wherein said learning result is probabilistic density distribution.

4. The behavior control apparatus claimed in claims 1 to 3, wherein said target segregation method segregates said portion by comparing the sensory inputs and said estimated motion.

5. The behavior control apparatus claimed in claim 4, wherein said segregation is done by utilizing optical flow.

6. The behavior control apparatus claimed in claims 1 to 5, wherein said target location acquiring method defines the center of the target object as the location of said target object;

said behavior decision method outputs the behavior command to move the mobile unit toward said location of the target object.

7. The behavior control apparatus claimed in claim 6, wherein said behavior decision method calculates the distance between the mobile unit and the location of said target object, said behavior decision method deciding the behavior command to decrease the calculated distance.

8. The behavior control apparatus claimed in claim 7, wherein if the calculated distance is greater than a predetermined value, said target segregation method repeats segregating said portion which includes a target object.

9. The behavior control apparatus claimed in claims 1 to 8, wherein said sensory input capturing method captures images of the external environment of the mobile unit as the sensory inputs.

10. The behavior control apparatus claimed in claims 1 to 8, wherein said target object matching method extracts target object by pattern matching between the sensory inputs and predetermined templates.

11. The behavior control apparatus claimed in claims 1 to 8, wherein said sensory inputs capturing method is a gyroscope which captures motion of the mobile unit.

12. A method for controlling behavior of a mobile unit using behavior command, comprising:

capturing sensory inputs;

estimating motion of the mobile unit;

segregating the portion which includes a target object to be target for behavior of the mobile unit from sensory inputs;

extracting the target object from said segregated portion;

acquiring the location of said target object; and

controlling the mobile unit based on the location of target object.

13. The method claimed in claim 12, said estimating step further comprising:

outputting said behavior command;

evaluating the result of the behavior of the mobile unit;

learning the motion of the mobile unit using the relationship between said sensory inputs and said behavior result; and

storing the learning result.

14. The method claimed in claim 13, wherein said learning result is probabilistic density distribution.

15. The method claimed in claims 12 to 14, wherein said segregating is done by comparing the sensory inputs and said estimated motion.

16. The method claimed in claim 15, wherein said segregation is done by utilizing optical flow.

17. The method claimed in claims 12 to 16, wherein the center of the target object is defined as the location of said target object;

said behavior command being determined so as to move the mobile unit toward said location of the target object.

18. The method claimed in claim 17, wherein the distance between the center of motion of the mobile unit and the location of said target object is calculated, and then the behavior command is determined to decrease the calculated distance.

19. The method claimed in claim 18, wherein if the calculated distance is greater than a predetermined value, said segregating step is repeated.

20. The method claimed in claims 12 to 19, wherein said sensory inputs are images of the external environment of the mobile unit.

21. The method claimed in claims 12 to 19, wherein said extracting is done by pattern matching between the sensory inputs and predetermined templates.

22. The method claimed in claim 12 to 19, wherein motion of the mobile unit is captured using a gyroscope.

23. Computer software program for implementing any of the method claimed in claims 12 to 22 on a computer.

24. A recording medium for recording said computer software program claimed in claim 23.