KR102257746B1 - Method for controlling robot group and system thereof - Google Patents

Abstract

The present invention relates to a robot group control method and system, and more specifically, while a management robot runs, generates map information for a work area, divides one or more sub-work areas from the work area, and assigns it to one or more work robot Thus, it relates to a robot group control method and system for controlling each work robot to perform a given task in each sub work area.
In the present invention, a map information generation step of generating map information on a work area while the management robot is traveling; A sub-work area allocating step of dividing one or more sub-work areas in the work area and assigning them to one or more work robots; And a task performing step of controlling the one or more work robots to perform respective tasks in the one or more sub-work areas.

Description

Robot group control method and system TECHNICAL FIELD

The present invention relates to a robot group control method and system, and more specifically, while a management robot travels, it generates map information for a work area, divides one or more sub-work areas from the work area, and assigns it to one or more work robots. Thus, it relates to a robot group control method and system for controlling each work robot to perform a given task in each sub work area.

With the popularization of robots, various services using robots have been provided in various fields, and for more specific examples, attempts have been made to provide information services using robots in airports or large shopping malls.

However, in the case of performing work in a large space such as an airport or a large shopping mall, the work can be performed by dividing and allocating the space using a plurality of robots. However, for this, the manager must allocate the work areas for the plurality of robots individually. There was a hassle.

In addition, in order to perform a task while the robot moves, map information on the work area is required, so that the robot generates map information through autonomous driving, but in airports or large shopping malls, areas having various environments may exist. In particular, even in a situation where it is difficult to properly secure lighting necessary for spatial recognition, such as in a dark space, it is difficult to generate accurate map information and divide and allocate the work area.

Furthermore, when a plurality of robots perform tasks while moving within the work area as above, expensive sensors and autonomous driving devices for autonomous driving must be installed for each of the plurality of robots, and thus the complexity of the robot and system and the manufacturing cost accordingly. There will be a problem that is greatly increased.

Therefore, the manager can perform tasks by dividing and assigning work areas for each of the plurality of robots without directly allocating the work areas to the plurality of robots. Also, it is possible to generate accurate map information even in environments where appropriate lighting is not secured. In addition, there is a request for a method that can be allocated by dividing and inhibiting the increase in the complexity of the robot and the system and the increase in manufacturing cost accordingly, but an appropriate solution has not yet been presented.

Republic of Korea Patent Publication No. 10-2018-0039438 (published on April 18, 2018)

The present invention was invented to solve the problems of the prior art as described above, and an administrator can perform a task by dividing and assigning a work area for each of a plurality of robots without directly allocating a work area to a plurality of robots. Robot control method and system capable of dividing and allocating work areas by generating accurate map information even in environments where proper lighting is not secured, and further suppressing the increase in the complexity of the robot and the system and the increase in manufacturing cost. It aims to provide.

In addition, detailed objects of the present invention will be clearly understood and understood by experts or researchers in this technical field through the detailed contents described below.

A method for controlling a robot group according to an aspect of the present invention for solving the above problem includes: a map information generating step of generating map information for a work area while a management robot is traveling; Sub-work area allocation step of dividing one or more sub-work areas in the work area and assigning them to one or more work robots; And a task performing step of controlling the one or more work robots to perform respective tasks in the one or more sub-work areas.

At this time, in the map information generation step, the management robot generates the map information by having an object detection and distance measurement sensor, and in the task execution step, the work robot detects an object and detects an object using the generated map information. It is characterized in that the operation is performed while traveling in the sub-work area without using a distance measurement sensor.

In addition, in the task performing step, the working robot generates image information about the surroundings by having an image camera, and calculates its own position information in the work area using the image information and the map information to perform a task. You can do it.

Furthermore, the management robot may control the location information of the work robot using relative location information of the work robot calculated based on its own location.

In addition, in the sub-work area allocation step, among the 3D point clouds collected for a specific area of the work area, a departure point group that deviates from the wall surface of the work space is calculated, and the number of the departure point groups is calculated. It can be determined whether or not a specific area is to be calculated as a sub-work area.

Here, the sub-work area allocating step includes: calculating a plane (2D) vector for a wall surface of the specific area by projecting a 3D point cloud collected for a specific area of the work area; Calculating a wall surface of the specific area from the plane (2D) vector; And calculating a departure point group deviating from the wall surface of the specific region among the three-dimensional point groups.

In addition, in the sub-work area allocation step, if the number of deviating point groups is less than or equal to a predetermined noise reference value, the deviating point group is determined as noise, and when the number of deviating point groups is greater than a predetermined noise reference value, the The deviated point group can be determined as an object.

In addition, in the task performing step, the task robot may perform a task by recognizing the sub-work area using an infrared image generated using an infrared sensor.

Further, the step of performing the work may include: (a) normalizing, by the work robot, a histogram of the infrared image; (b) applying a morphology operation-erosion-and Gaussian filtering-blur-to the infrared image; (c) binarizing the infrared image; And (d) performing blob labeling on the infrared image.

In addition, in the map information generation step, the management robot includes an image camera and an object detection and distance measurement sensor, and when the illuminance determined through the image data of the image camera does not meet a predetermined illuminance reference value, the object is detected. And it is possible to generate the map information using a distance measurement sensor.

In a robot group control system according to another aspect of the present invention, in a control system for a robot group including a management robot and one or more work robots, map information on a work area is generated while driving, and at least one A management robot that divides the sub work area and assigns it to one or more work robots; And a work robot that performs work in the assigned one or more sub work areas.

At this time, the management robot is equipped with an object detection and distance measurement sensor to generate the map information, and the work robot uses the generated map information to drive the sub-work area without using an object detection and distance measurement sensor. And can do the job.

Accordingly, in the method and system for controlling a robot group according to an embodiment of the present invention, more specifically, while the management robot is traveling, map information on a work area is generated, and at least one sub work area is divided from the work area. By assigning to one or more work robots and controlling each work robot to perform a given task in each sub work area, the work area is divided and assigned to each of the plurality of robots without the administrator directly allocating the work area to the plurality of robots. You will be able to do it.

In addition, in the robot group control method and system according to an embodiment of the present invention, even in an environment where the robot does not have adequate lighting, it is possible to divide and allocate the work area by generating accurate map information, and furthermore, the complexity of the robot and the system It is possible to effectively suppress the increase and thus the increase in the manufacturing cost.

The accompanying drawings, which are included as part of the detailed description to aid in understanding of the present invention, provide embodiments for the present invention, and describe the technical spirit of the present invention together with the detailed description.
1 is a flowchart of a method for controlling a robot group according to an embodiment of the present invention.
2 is a block diagram of a robot group control system according to an embodiment of the present invention.
3 is a diagram illustrating an operation of a method and a system for controlling a robot group according to an embodiment of the present invention.
4 is a diagram illustrating information transmission between a management robot and a work robot in a robot group control method and system according to an embodiment of the present invention.
5 to 9 are diagrams for explaining calculation of a sub-work area in a method and system for controlling a robot group according to an embodiment of the present invention.
10 is a diagram illustrating a method of controlling a robot group and a method of controlling a position of a working robot based on a position of a management robot in a system and a method of controlling a robot group according to an embodiment of the present invention.
11 is a view for explaining the operation of a management robot and a work robot in the robot group control method and system according to an embodiment of the present invention.
12 is a diagram illustrating an algorithm for calculating an object from an infrared image in a robot group control method and system according to an embodiment of the present invention.

The present invention may apply various transformations and may have various embodiments. Hereinafter, specific embodiments will be described in detail based on the accompanying drawings.

The following examples are provided to aid in a comprehensive understanding of the methods, apparatus and/or systems described herein. However, this is only an example and the present invention is not limited thereto.

In describing the embodiments of the present invention, when it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, a detailed description thereof will be omitted. In addition, terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of users or operators. Therefore, the definition should be made based on the contents throughout the present specification. The terms used in the detailed description are only for describing the embodiments of the present invention, and should not be limiting. Unless explicitly used otherwise, expressions in the singular form include the meaning of the plural form. In the present description, expressions such as "comprising" or "feature" are intended to refer to certain features, numbers, steps, actions, elements, some or combination thereof, and one or more It should not be construed to exclude the presence or possibility of other features, numbers, steps, actions, elements, any part or combination thereof.

In addition, terms such as first and second may be used to describe various components, but the components are not limited by the terms, and the terms are used to distinguish one component from other components. Is only used.

Hereinafter, exemplary embodiments of a method and system for controlling a robot group according to an embodiment of the present invention will be described in order with reference to the accompanying drawings.

First, in FIG. 1, a flow chart of a method for controlling a robot group according to an embodiment of the present invention is shown. As can be seen in FIG. 1, the robot group control method according to an embodiment of the present invention includes a map information generation step (S100) of generating map information for a work area while the management robot 100 is traveling, the operation Sub-work area allocation step (S200) of dividing one or more sub-work areas in an area and allocating them to one or more work robots 200, and so that the one or more work robots 200 perform tasks in the one or more sub-work areas, respectively. It includes a controlling operation performing step (S300).

At this time, the management robot 100 is provided with an object detection and distance measurement sensor capable of detecting a nearby object such as a LiDAR, a radar depth sensor, and measuring the distance to the object. Map information for the work area can be generated, and accordingly, the work robot 200 uses the map information generated by the management robot 100 to detect an object and use a distance measurement sensor in the work area. You can perform tasks while driving through the divided sub-work areas.

Accordingly, in the robot group control method according to an embodiment of the present invention, the management robot 100 is equipped with the object detection and distance measurement sensor to autonomously travel while the task robot 200 can perform a task. Map information for an area can be generated, and each work robot 200 performs a given task in each sub work area by dividing one or more sub work areas in the work area and assigning it to one or more work robots 200. By controlling so that the manager does not directly allocate the work area to the plurality of robots 200, the work area is divided and assigned to each of the plurality of robots to perform the task.

Further, in the robot group control method according to an embodiment of the present invention, the work robot 200 uses the map information generated by the management robot 100 without using an object detection and distance measurement sensor to determine the sub work area. By performing work while driving, it is possible to effectively suppress an increase in the complexity of robots and systems and the resulting increase in manufacturing cost.

In addition, FIG. 2 shows a configuration diagram of a robot group control system 10 according to an embodiment of the present invention. As can be seen in FIG. 2, the robot group control system 10 according to an embodiment of the present invention is a control system for a robot group including a management robot 100 and one or more work robots 200, A management robot 100 that generates map information for a work area while driving, divides one or more sub-work areas from the work area, and assigns it to one or more work robots, and performs a task in the assigned one or more sub-work areas. It may be configured to include a working robot 200.

At this time, the management robot 100 may generate map information on the work area by having an object detection and distance measurement sensor, and accordingly, the work robot 200 detects an object using the generated map information. And without using a distance measurement sensor, it is possible to perform a task while traveling through the sub-work area divided in the work area.

Hereinafter, a robot group control method and system 10 according to an embodiment of the present invention will be divided for each configuration and examined in more detail with reference to FIGS. 1 and 2.

First, in the map information generating step (S100), the management robot 100 generates map information for a work area while driving.

At this time, the management robot 100 is provided with an object detection and distance measurement sensor capable of detecting a nearby object such as a LiDAR, a radar depth sensor, and measuring the distance to the object. It is possible to generate map information for the work area while driving autonomously. Subsequently, the management robot 200 may transmit the map information on the generated work area to one or more work robots 200, and further, by dividing one or more sub work areas in the work area, one or more work robots ( 200).

Accordingly, the work robot 200 uses the map information generated by the management robot 100 to drive the sub work area divided and allocated in the work area without using an object detection and distance measurement sensor. You will be able to do it. Furthermore, through this, in the robot group control method and system according to an embodiment of the present invention, even if only the management robot 100 is provided with an object detection and distance measurement sensor, one or more work robots 200 may be used in the management robot 100. Since it is possible to perform a task while driving within the work area using the generated map information, it is possible to effectively suppress an increase in the complexity of the robot and system and the resulting increase in manufacturing cost.

That is, as can be seen in FIG. 3, in the prior art, the working robot 200, which does not have an object detection and distance measurement sensor, has a problem in that it cannot detect the surrounding environment and thus cannot perform a task while traveling by itself (FIG. 3 (a)), in the robot group control method and system according to an embodiment of the present invention, while the management robot 100 having an object detection and distance measurement sensor travels, map information on a work area is generated, and the work robot 200 ), and further divides the work area into one or more sub work areas and allocates it to the work robot 200 ((A), (B), (C) in Fig. 3(b)) and the transmitted map information The sub-work area is used to perform a task (FIG. 3(b)).

More specifically, FIG. 4 illustrates information transmission between the management robot 100 and the work robot 200 in the robot group control method and system 10 according to an embodiment of the present invention. As can be seen in FIG. 4, the management robot 100 generates and shares map information on the work area, further allocates a sub work area to the work robot 200 and transmits a control command. Accordingly, the work robot 200 performs work in the assigned sub work area, and further transmits its own location information or location information of a customer detected through image analysis, etc., to the management robot 100 It can also be used for control.

Further, in the robot group control method and system 10 according to an embodiment of the present invention, the management robot 100 includes an image camera and an object detection and distance measurement sensor, and is determined through image data of the image camera. When the illumination intensity does not meet a predetermined illumination reference value, the map information is generated using the object detection and distance measurement sensor, so that map information having high reliability can be generated even in an environment in which illumination is insufficient.

Subsequently, in the sub-work area allocation step (S200), the management robot 100 divides one or more sub-work areas from the work area and assigns them to the one or more work robots 200.

At this time, in the sub-work area allocation step (S200), the management robot 100 calculates a deviated point group deviating from the wall surface of the work space among 3D point clouds collected for a specific area of the work area. And, it may be determined whether to calculate the specific area as a sub-work area according to the number of the departure point group.

More specifically, as can be seen in FIG. 5, in the sub-work area allocation step (S200), a 3D point cloud collected for a specific area of the work area is projected onto the wall surface of the specific area. Calculating a plane (2D) vector for (Fig. 5(D)), calculating a wall surface of the specific area from the plane (2D) vector (Fig. 5(E)), and the specific among the three-dimensional point groups It may include a step ((F) of FIG. 5) of calculating a departure point group that deviates from the wall surface of the region.

)로부터 가상의 선을 그어 상기 벽면의 라인을 산출한다. Accordingly, the management robot 100, as can be seen in Figure 6, using the map information generated in the map information generation step (S100), an image camera, an infrared sensor or a Kinect sensor, etc. A plane (2D) vector representing a wall surface in a specific area of the work area calculated from the generated 3D point cloud (Fig.

) To calculate the line of the wall by drawing an imaginary line.

Subsequently, the points deviating from the calculated wall line may be objects such as customers located in a work space such as a hall or a room, or may be determined as noise. In this case, as shown in FIG. 5, when the deviating points exceed a predetermined first reference value Tw, the area where the points are located is determined as a large work area capable of searching and responding to objects such as customers, and If the deviating points are smaller than the first reference value Tw and larger than the second reference value Te, it is determined as a small work area such as a room smaller than the large work area, and is smaller than the second reference value Te. In this case, it can be judged as noise. In this case, it is preferable to calculate the first reference value Tw and the second reference value Te in consideration of the type and characteristic of the sensor to be used.

Accordingly, in the robot group control method and system 10 according to an embodiment of the present invention, a dynamic work space capable of performing work by reflecting a work environment that changes frequently, not a static work space, is secured to perform work. You can do it. At this time, the management robot 100 determines and secures an area in which the work robot 200 can perform the work while moving smoothly because there are no obstacles or the like that interfere with the work performance while driving dynamically.

In the robot group control method and system 10 according to an embodiment of the present invention, the management robot 100 is a 3D point cloud generated using an image camera, an infrared sensor or a kinect sensor. ), the plane is calculated using the plane model equation of the point group. At this time, in the robot group control method memetic system according to an embodiment of the present invention, it is preferable to apply an algorithm capable of effectively deriving model parameters even from original data with high noise, such as the RANSAC algorithm. More specifically, the RANSAC algorithm derives an optimal solution by iteratively calculating while randomly sampling minimum data necessary to determine a parameter of a model among all original data. Conventional statistical methods use as much data as possible to obtain an initial solution and remove invalid data from the result, whereas the RANSAC algorithm increases a consistent set of data by using as little initial data as possible. You will use the method.

Accordingly, in applying the RANSAC algorithm, the number of repetitions N must be determined, and it is preferable to select a sufficiently large N so as to guarantee a predetermined probability reference value (typically 99%). Here, if the probability that the data is valid is u, the probability that the data is not valid is v = 1-u, and at this time, the number of repetitions N for the number of sample data m can be calculated by Equation 1 below.

[Equation 1]

In this regard, FIG. 7 illustrates a result of recognizing a plane followed by a point cloud by performing RANSAC segmentation using a plane model in a three-dimensional space.

More specifically, in the robot group control method and system 10 according to an embodiment of the present invention, a process of calculating a plane followed by a point group may be divided into two steps. First, in the first step, as shown in FIG. 8, a normal vector is calculated from the point group, and whether the calculated normal vector is in the same plane is determined based on the angular threshold (for example, As an embodiment of the present invention, 3° can be used as the angular threshold value). Accordingly, the average vector [A, B, C] ^T can be obtained using the normal vector of points located in the same plane, and accordingly, the plane can be expressed by Equation 2 as follows, where d is [ x _m , y _m , z _m ] ^T can be utilized.

[Equation 2]

을 구하고, 이를 중력 벡터

와의 내적을 통해 아래 수학식 3과 같이 중력 벡터와의 각도를 구한다. Subsequently, in the second step, the accuracy of the estimated plane is improved as shown in FIG. 9. At this time, the floor plane on which the robot travels and performs tasks is a stable space without obstacles and becomes the robot's work space. Accordingly, the representative normal vector of each plane is used to distinguish the left, right and floor planes.

Find the gravitational vector

Through the dot product of and, the angle with the gravity vector is obtained as in Equation 3 below.

[Equation 3]

In addition, as can be seen in FIG. 9, the left, right, and floor planes are sequentially calculated. Since the robots are located parallel to the floor plane, the initial value of θ is 180°, and if the value of θ is -90° or 270°, it becomes the right plane, and if it is 90°, it can be estimated as the left plane.

가 도출될 수 있으며, 이어서 상기 평면(2D) 벡터

을 이용하여 서브 작업 영역을 산출하게 된다. 이에 따라, 본 발명에서는 영상의 단일 이미지로부터 상기 벽면 벡터를 구하여 벽면을 추정할 수 있어, 종래 기술과 같이 연속적 이미지의 분석을 통해 벽면을 추정하는 경우보다 적은 연산량으로 빠르게 처리할 수 있다는 장점을 가진다.That is, in the robot group control method and system 10 according to an embodiment of the present invention, in order to detect a plane, a point group is divided using the normal vector of each point, and the floor, right, and left planes are estimated from the result. do. At this time, as can be seen in Figure 6, from P _{1 of one end and P 2} of the other end

Can be derived, and then the plane (2D) vector

The sub-work area is calculated using. Accordingly, in the present invention, the wall surface can be estimated by obtaining the wall surface vector from a single image of an image, and thus has the advantage that it can be processed faster with less computational amount than when estimating the wall surface through continuous image analysis as in the prior art. .

Subsequently, in the task execution step S300, the management robot 100 controls the one or more task robots 200 to perform tasks in the one or more sub-work areas, respectively.

At this time, the working robot 200 is equipped with an image camera to generate image information about the surroundings, and using the image information and the map information transmitted from the management robot 100, its own position information in the work area. Can be calculated, and accordingly, it is possible to perform work while freely traveling within the work area without having an expensive sensor such as LiDAR.

Furthermore, at this time, as shown in FIG. 10, the management robot 100 uses the relative position information of the work robot 200 calculated based on its own position to provide the position information of the work robot 200. Can be controlled.

That is, in the robot group control method and system 10 according to an embodiment of the present invention, map information is created based on the management robot 100, and accordingly, the working robot 200 is the management robot 100 Location information may be defined based on the relative location of the family. The coordinate system setting accordingly may be defined as in Equation 4 below.

[Equation 4]

에 의해 위 수학식 4와 같이 표현될 수 있다. 여기서

는 Z_s 축에 대해 θ_s 만큼 회전한 후, 회전된 작업 로봇(200) 1의 좌표계의 X_T1축 방향으로 ^sP_T1ORG 만큼 이동하는 변환 행렬이다. 또한, 관리 로봇(100)은 작업 로봇(200) 2와의 좌표 변환 행렬

의 관계를 이용하여 관리 로봇(100)을 기준으로 작업 로봇(200) 2의 상대 위치 및 방위도 기술할 수 있게 된다.That is, as can be seen in FIG. 10, the relationship between the ^{point s} P based on the coordinate system {S} of the management robot 100 ^{and the point T1} P based on the coordinate system {T1} of the work robot 200 is Homogeneous transformation matrix between {S} and {T1}

It can be expressed as in Equation 4 above. here

Is a transformation matrix that _{is rotated by θ s} with respect to the Z _s ^{axis and then moved by s} P _{T1ORG in the X T1} axis direction of the rotated work robot 200 1. In addition, the management robot 100 is a coordinate transformation matrix with the work robot 200

It is possible to describe the relative position and orientation of the working robot 200 2 based on the management robot 100 using the relationship of.

Accordingly, in the robot group control method and system 100 according to an embodiment of the present invention, as shown in FIG. 11, the management robot 100 and the work robot 200 are divided into a wide work area through division of labor. A dynamic work space is secured for, and sub-work areas are divided and allocated so that one or more work robots 200 can smoothly perform work in each sub work area.

More specifically, the management robot 100 generates map information on the work area (FIG. 11(G)), and selects a sub work area in which the work robot 200 needs to be placed in the work area (Fig. 11 (H)). In addition, the management robot 100 secures a dynamic work area while autonomously traveling and provides it to the work robot 200 (FIG. 11(I)), so that the work robot 200 safely travels in the dynamic work area. It will be possible to perform tasks such as customer response smoothly. Accordingly, one or more work robots 200 are sequentially moved to the calculated sub-work areas to perform tasks such as searching and responding to customers.

At this time, the work robot 200 can recognize a customer, etc. based on an image and perform a task (Fig. 11 (J)), and when the work robot 200 starts a task such as customer response, the location of the customer And information such as a work situation may be transmitted to the management robot 100 or a management server. When there is no customer in the sub work area, the work robot 200 moves to another sub work area and repeats the process of performing the work. In this case, when no additional sub-work area remains, the management robot 100 moves to another area, generates map information for the additional area, secures the work area, and allocates it to the work robot 200.

In addition, in the robot group control method and system 10 according to an embodiment of the present invention, even when illuminance for capturing a camera image is not secured, a sub-work area is recognized using an infrared sensor, etc., and objects such as customers are You can navigate to and do the job.

More specifically, as can be seen in FIG. 12, the operation performing step (S300) includes (a) normalizing the histogram of the infrared image by the working robot (S1310), (b) the infrared image Applying a morphology operation-erosion-and Gaussian filtering-blur-to (S1320, S1330), (c) binarizing the infrared image (S1340) and (d) the infrared Including the step of performing blob labeling on the image (S1350), a sub-work area may be recognized using an infrared image and an object such as a customer may be searched to perform a task.

That is, FIG. 12 illustrates a process of recognizing and segmenting a sub-work area through an infrared image. In this case, the histogram of the infrared image may be defined as in Equation 5 below.

[Equation 5]

Here, X _k is the k-th brightness value, and n _k is the number of pixels with brightness X _{k in the image.} The histogram of the original infrared image is normalized by dividing each element of the histogram by the total number of pixels in the image because most of the pixels are concentrated in the dark area.

Accordingly, the normalized histogram p(X _k ) represents the probability of occurrence of brightness X _k , and an equation for obtaining the probability density function (PDF) may be defined as in Equation 6 below.

[Equation 6]

Here, n is the total number of pixels in the image, and the sum of all elements of the normalized histogram is 1. In addition, if the cumulative distribution function (CDF) is obtained and multiplied by the maximum contrast value (L-1), an image value to which the histogram smoothing is applied can be obtained as shown in Equation 7 below.

[Equation 7]

Here, s _k is a smoothed brightness value corresponding to the brightness value X _k of the image, and when applied to the entire image, the smoothed image has a clear contrast compared to the original infrared image.

Next, the background and the person are separated, and morphological operations such as erosion and Gaussian filter are applied to remove noise. First, the erosion operation may be defined as in Equation 8 below.

[Equation 8]

Here, when B is moved by z, if the moved B is completely a subset of A, it becomes B. In FIG. 12(K), an infrared image to which an erosion operation is applied is exemplified.

Next, a Gaussian filter, which is a low frequency filter, may be defined as in Equation 9 below.

[Equation 9]

Where (x, y) is the position of the pixel in the image and σ is the standard deviation. After applying binarization in the image from which noise has been removed (FIG. 12(L)), 4-Connectivity Blob Labeling is performed to recognize objects such as customers. In the labeling technique, in a binary image divided into a background and an object, four neighboring pixels are identified, and an object is labeled by assigning the same ID as the starting point of each object. If labeling of all objects is applied in this way, several labeled objects are recognized. Among them, objects such as customers are recognized by classifying noise and people according to the number of pixels in the labeling (FIG. 12(M)). Through such a series of processes, in the robot group control method and system 10 according to an embodiment of the present invention, the sub-work area is recognized and objects such as customers are searched for work even in a space where illumination or light does not come on. Will be able to perform smoothly.

The above description is merely illustrative of the technical idea of the present invention, and those of ordinary skill in the art to which the present invention pertains will be able to make various modifications and variations without departing from the essential characteristics of the present invention. Therefore, the embodiments described in the present invention are not intended to limit the technical idea of the present invention, but to explain, and are not limited to these embodiments. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

10: Robot group control system
100: management robot
200: working robot

Claims (12)

A map information generation step of generating map information on a work area while the management robot travels;
A sub-work area allocating step of dividing one or more sub-work areas in the work area and assigning them to one or more work robots; And
A task performing step of controlling the one or more work robots to perform tasks in the one or more sub-work areas, respectively;
Including,
The sub-work area allocation step,
Calculate a departure point group deviating from the wall surface of the work space among the 3D point clouds collected for a specific area of the work area,
And determining whether to calculate the specific area as a sub-work area according to the number of the deviated point groups. The method of claim 1,
In the step of generating the map information,
The management robot is provided with an object detection and distance measurement sensor to generate the map information,
In the step of performing the above operation,
The robot group control method, wherein the working robot performs a task while traveling through the sub-work area without using an object detection and distance measurement sensor using the generated map information. The method of claim 2,
In the step of performing the above operation,
The working robot has an image camera to generate image information about the surroundings,
A method of controlling a robot group, characterized in that performing a task by calculating its own location information in the work area using the image information and the map information. The method of claim 3,
The management robot,
The robot group control method, characterized in that the position information of the working robot is controlled by using the relative position information of the working robot calculated based on its own position. delete The method of claim 1,
The sub-work area allocation step,
Calculating a plane (2D) vector for a wall surface of the specific area by projecting a 3D point cloud collected for a specific area of the working area;
Calculating a wall surface of the specific area from the plane (2D) vector; And
And calculating a departure point group that deviates from the wall surface of the specific area among the three-dimensional point groups. The method of claim 1,
In the sub-work area allocation step,
When the number of the deviated point groups is less than or equal to a predetermined noise reference value, the deviating point group is determined as noise,
When the number of the deviated point groups is greater than a predetermined noise reference value, the deviating point group is determined as an object. The method of claim 1,
In the step of performing the above operation,
The robot group control method, wherein the working robot performs a task by recognizing the sub-work area using an infrared image generated using an infrared sensor. The method of claim 8,
The step of performing the operation,
(a) normalizing, by the working robot, a histogram of the infrared image;
(b) applying a morphology operation-erosion-and Gaussian filtering-blur-to the infrared image;
(c) binarizing the infrared image; And
(d) performing blob labeling on the infrared image. The method of claim 1,
In the step of generating the map information,
The management robot includes an image camera and an object detection and distance measurement sensor,
When the illuminance determined through the image data of the video camera does not meet a predetermined illuminance reference value,
And generating the map information using the object detection and distance measurement sensor. In a control system for a robot group comprising a management robot and one or more work robots,
A management robot that generates map information on a work area while driving, divides one or more sub-work areas from the work area, and assigns it to one or more work robots; And
A work robot that performs work in the assigned one or more sub work areas;
Including,
The management robot,
Calculate a departure point group deviating from the wall surface of the work space among the 3D point clouds collected for a specific area of the work area,
And determining whether to divide and allocate the specific area into sub-work areas according to the number of the deviated point groups. The method of claim 11,
The management robot is provided with an object detection and distance measurement sensor to generate the map information,
The robot group control system, wherein the working robot performs a task while traveling through the sub-work area without using an object detection and distance measurement sensor using the generated map information.

Images