ROBOT CLEANER AND CONTROL METHOD THEREOF

• US 20160089783A1
• Filed: 09/29/2015
• Published: 03/31/2016
• Est. Priority Date: 09/30/2014
• Status: Active Grant

• Current Assignee: LG Electronics, Inc. (LG Corporation)
• Original Assignee: LG Electronics, Inc. (LG Corporation)
• Inventors: Noh, Dongki, Lim, Seungwook, Lee, Taekyeong, Shim, Inbo
• Granted Patent: US 9,798,957 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G05D 1/0225 involving docking at a fixed facility, e.g. base station or loading bay parking aids B62D15/027
• CPC Class Codes: G05D 1/0246 using a video camera in combination with image processing means
• CPC Class Codes: G05D 1/0274 using mapping information stored in a memory device navigation using map-matching G01C21/30
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06F 18/24765 Rule-based classification
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/467 Encoded features or binary features, e.g. local binary patterns [LBP]
• CPC Class Codes: G06V 10/758 Involving statistics of pixels or of feature values, e.g. histogram matching
• CPC Class Codes: G06V 10/765 using rules for classification or partitioning the feature space
• CPC Class Codes: G06V 20/20 in augmented reality scenes
• CPC Class Codes: Y10S 901/01 Mobile robot

Abstract

A control method for a robot cleaner includes acquiring a plurality of images of surroundings during travel of the robot cleaner in a cleaning area, estimating a plurality of room-specific feature distributions according to a rule defined for each of a plurality of rooms, based on the images acquired while acquiring the plurality of images, acquiring an image of surroundings at a current position of the robot cleaner, obtaining a comparison reference group including a plurality of room feature distributions by applying the rule for each of the plurality of rooms to the image acquired while acquiring the image at the current position, comparing the obtained comparison reference group with the estimated room-specific feature distributions, and determining a room from the plurality of rooms having the robot cleaner currently located therein.

Claims

1. A control method for a robot cleaner, comprising:

acquiring a plurality of images of surroundings during travel of the robot cleaner in a cleaning area, by an image acquisition unit;

estimating a plurality of room-specific feature distributions according to a rule defined for each of a plurality of rooms, based on the images acquired while acquiring the plurality of images, by a feature distribution learning module;

acquiring an image of surroundings at a current position of the robot cleaner, by the image acquisition unit;

obtaining a comparison reference group comprising a plurality of room feature distributions by applying the rule for each of the plurality of rooms to the image acquired while acquiring the image at the current position, by a position recognition module;

comparing the obtained comparison reference group with the estimated room-specific feature distributions, by the position recognition module; and

determining a room from the plurality of rooms having the robot cleaner currently located therein, by the position recognition module.

2. The control method according to claim 1, wherein the estimating a plurality of room-specific feature distributions comprises:

estimating a plurality of room feature distributions for the images acquired in the acquiring of the plurality of images; and

estimating a room-specific feature distribution for each of the plurality of rooms from the images acquired based on the estimated feature distributions for the images.

3. The control method according to claim 2, wherein the estimating of the plurality of room feature distributions comprises:

detecting a plurality of room features in the images acquired in the acquiring of the plurality of images;

creating a plurality of labels for the room features according to the rule defined for each of the rooms; and

scoring the labels and estimating a label-specific score distribution for each of the images;

wherein the estimating of the room-specific feature distribution comprises;

estimating a label-specific score distribution for each of the rooms based on the label-specific score distribution for each of the images.

4. The control method according to claim 3, wherein the scoring of the labels is performed according to the rule defined for each of the rooms.

5. The control method according to claim 3, wherein the estimated label-specific score distribution for each of the rooms is obtained by averaging the label-specific score distributions for the images according to each of the rooms.

6. The control method according to claim 2, further comprising:

comparing a feature distribution subjected to the rule for the room having the robot cleaner currently located therein among the feature distributions of the comparison reference group with the feature distributions for the images; and

determining a region having the robot cleaner positioned therein in the room.

7. The control method according to claim 1, further comprising:

recognizing a position of the robot cleaner based on a pre-stored map of the cleaning area,

wherein the acquiring an image of surroundings, the obtaining a comparison reference group, and the comparing the obtained comparison reference group with the estimated the room-specific feature distribution are performed when the position of the robot cleaner on the map is not recognized.

8. The control method according to claim 7, wherein the map is input through an external terminal, the external terminal able to wirelessly communicate with the robot cleaner.

9. The control method according to claim 1, wherein the images of the surroundings are acquired by photographing a ceiling within the cleaning area.

10. The control method according to claim 1, wherein the rules for each of the rooms are different from each other.

11. A robot cleaner, comprising:

an image acquisition unit to acquire a plurality of images of surroundings of the robot cleaner;

a feature distribution learning module to estimate a plurality of room-specific feature distributions according to a rule defined for each of a plurality of rooms, based on the images acquired through the image acquisition unit; and

a position recognition module to obtain a comparison reference group comprising a plurality of room feature distributions by applying the rule for each of the rooms to an image acquired at a current position of the robot cleaner through the image acquisition unit and to compare the comparison reference group with the room-specific feature distributions estimated through the feature distribution learning module and determine a room from the plurality of rooms having the robot cleaner currently located therein.

12. The robot cleaner according to claim 11, wherein the feature distribution learning module estimates a room-specific feature distribution for each of the rooms used as the surroundings for the acquired images based on the room feature distributions for the images.

13. The robot cleaner according to claim 11, further comprising a feature detection module to detect a plurality of room features in the images acquired by the image acquisition unit,

wherein the feature distribution learning module creates, according to the rule defined for each of the rooms, a plurality of labels for the room features detected by the feature detection module, estimates a label-specific score distribution for each of the images by scoring the labels, and estimates a label-specific score distribution for each of the rooms based on the label-specific score distribution for each of the images.

14. The robot cleaner according to claim 13, wherein the labels are scored according to the rule defined for each of the rooms.

15. The robot cleaner according to claim 13, wherein the estimated label-specific score distribution for each of the rooms is obtained by averaging the label-specific score distributions for the images according to each of the rooms.

16. The robot cleaner according to claim 12, wherein the position recognition module compares a feature distribution subjected to the rule for the room determined to have the robot cleaner currently located therein among the feature distributions of the comparison reference group with the feature distributions for the images acquired in the room determined to have the robot cleaner located therein, and determines a region having the robot cleaner positioned therein in the room.

17. The robot cleaner according to claim 11, wherein the image acquisition unit acquires the images by photographing a ceiling of a cleaning area.

18. A control method for a robot cleaner, comprising:

acquiring a plurality of images of surroundings by moving the robot cleaner in a cleaning area, by an image acquisition unit;

detecting a plurality of room features in a plurality of images acquired for each of a plurality of rooms, creating a plurality of labels for the room features, scoring each of the labels, and estimating and storing a label-specific score distribution for each of the rooms, according to a rule defined for each of the rooms, by a feature distribution leaning module; and

acquiring an image of surroundings after the label-specific score distributions are stored, by an image acquisition unit;

creating a plurality of labels for a plurality of room features for the image according to the rule for each of the rooms, by the feature distribution leaning module;

obtaining a comparison reference group comprising a plurality of label-specific score distributions according to the rule defined for each of the rooms, by the feature distribution leaning module; and

determining a room having the robot cleaner currently located therein based on the comparison reference group and the label-specific score distribution estimated for each of the rooms, by the feature distribution leaning module.