KR102547274B1 - Moving robot and method for estiating location of moving robot - Google Patents

Abstract

The present application relates to a mobile robot and a method for recognizing a location thereof, and a method for recognizing a location of a mobile robot according to an embodiment of the present invention relates to a method for recognizing a location of a mobile robot including a camera and a laser sensor, wherein the camera and A matching step of receiving image information and depth information at time t from a laser sensor, respectively, and calibrating the received image information and depth information; Among the image information, a region irradiated with a laser line of the laser sensor is set as a region of interest (ROI), and the depth information is combined with an edge profile extracted from the region of interest to obtain the time t Characteristic information generation step of generating feature information of; and location estimation that compares feature information at time t and partial feature information at time t-1 using an Iterative Closest Point (ICP) algorithm, and estimates the pose and position of the mobile robot through the comparison result. steps may be included.

Description

Moving robot and method for recognizing its location {Moving robot and method for estiating location of moving robot}

The present application relates to a mobile robot and a method for recognizing a location thereof, and more particularly, to a mobile robot capable of estimating its own position using a camera and a laser sensor and a method for recognizing a location thereof.

Recently, with the development of robot technology, a mobile robot that sets a path and moves by itself is being used. In order for a mobile robot to effectively determine its position in space and move, it is required to recognize its own position in space while generating a map of the space it is moving in.

The mobile robot travels by dead reckoning using a gyroscope and an encoder provided in a drive motor, and analyzes an image using a camera installed on the top to create a map. At this time, driving information from the gyroscope and encoder If an error occurs due to , the accumulated error is corrected using the image information obtained from the camera.

Registered Patent Publication No. 10-1776621 (2017.09.11)

The present application is intended to provide a mobile robot capable of estimating its own position using a camera and a laser sensor and a method for recognizing its position.

A method for recognizing a location of a mobile robot according to an embodiment of the present invention relates to a method for recognizing a location of a mobile robot including a camera and a laser sensor, and receives image information and depth information at a point in time t from the camera and laser sensor, respectively. and a calibrating step of calibrating the received image information and depth information; Among the image information, a region irradiated with a laser line of the laser sensor is set as a region of interest (ROI), and the depth information is combined with an edge profile extracted from the region of interest to obtain the time t Characteristic information generation step of generating feature information of; and location estimation that compares feature information at time t and partial feature information at time t-1 using an Iterative Closest Point (ICP) algorithm, and estimates the pose and position of the mobile robot through the comparison result. steps may be included.

Here, in the matching step, coordinate conversion may be performed according to the installation positions of the camera and the laser sensor, and matching coordinates may be set so that the coordinates of the image information and the depth information match.

Here, in the feature information generating step, a region irradiated with the laser line may be extracted from the image information using the matching coordinates, and a range separated by a set interval based on the laser line may be set as the region of interest.

In the feature information generation step, a Sobel mask may be applied to the region of interest to extract an edge included in the region of interest, and the edge profile may be generated using the extracted edge. .

Here, the depth information may include a distance measurement value to each measurement point that equally divides the measurement angle range of the laser sensor into reference angle intervals.

Here, in the feature information generating step, the feature information may be generated by adding the depth information to coordinates corresponding to the measurement point in the edge profile.

Here, the partial feature information may include only feature information of measurement points included in a preset limit angle range among the measurement angle ranges of the laser sensor.

In the location estimating step, a point cloud model for the region of interest may be generated using the plurality of feature information.

Here, in the position estimation step, a target area corresponding to the partial feature information at time t-1 included in the point cloud model at time t is extracted, and the target area at time t-1 and the target at time t A pose and location of the mobile robot may be estimated by comparing locations of regions.

을 이용하여 N개의 측정포인트를 포함하는 상기 t 시점의 특징정보를 상기 포인트 클라우드 모델로 생성하고,

, Q>P, ∀P, Q ∈ [1,...,N]을 이용하여 상기 t-1 시점의 부분 특징정보에 대응하는 타겟 영역을 생성하며, 이때, 상기 측정각도 범위는 첫번째 측정포인트(i=1)로부터 N번째 측정포인트(i=N) 사이의 각도이고, 상기 제한각도 범위는 상기 P번째 측정포인트(i=P)로부터 Q번째 측정포인트(i=Q) 사이의 각도이며, 상기

,

이고, d⁽ⁱ⁾ _t-1은 t-1시점에서 i번째 측정포인트의 이동거리, Φ⁽ⁱ⁾ _t-1은 t-1 시점에서 i번째 측정포인트의 회전각, 각 측정포인트 사이의 각도차는 A일 수 있다. Here, the location estimation step,

Generating the feature information of the time point t including N measurement points into the point cloud model using

, Q>P, ∀P, Q ∈ [1,...,N] to create a target area corresponding to the partial feature information at the time point t-1. At this time, the measurement angle range is the first measurement point The angle between (i = 1) and the Nth measurement point (i = N), and the limit angle range is the angle between the P th measurement point (i = P) and the Q th measurement point (i = Q), remind

,

, d ⁽ⁱ⁾ _t-1 is the movement distance of the ith measurement point at time t-1, Φ ⁽ⁱ⁾ _t-1 is the rotation angle of the ith measurement point at time t-1, and the angle between each measurement point Car can be A.

Here, in the position estimation step, a limit angle range of the partial feature information may be set according to a maximum movement distance and a maximum rotation angle that the mobile robot can move in each frame.

에서, 상기 E_estimate가 최소값이 되는 이동거리 d 및 회전각 Φ를 추출하고, 상기 추출된 이동거리 d 및 회전각 Φ를 상기 최대이동거리 및 최대회전각으로 설정할 수 있다. Here, the location estimation step,

In , a movement distance d and rotation angle Φ at which the E _estimate is a minimum value may be extracted, and the extracted movement distance d and rotation angle Φ may be set as the maximum movement distance and maximum rotation angle.

According to an embodiment of the present invention, a computer program stored in a medium may exist in order to execute the method for recognizing the location of the mobile robot in combination with hardware.

A mobile robot according to an embodiment of the present invention includes a camera for generating image information; a laser sensor generating depth information; a matching unit for receiving image information and depth information at time t, and calibrating the received image information and depth information; Among the image information, the region irradiated with the laser line of the laser sensor is set as a region of interest (ROI), and the depth information is combined with an edge profile extracted from the region of interest to obtain the time point t Characteristic information generating unit for generating feature information of; and location estimation that compares feature information at time t and partial feature information at time t-1 using an Iterative Closest Point (ICP) algorithm, and estimates the pose and position of the mobile robot through the comparison result. wealth may be included.

In addition, the solution to the above problem does not enumerate all the features of the present invention. Various features of the present invention and the advantages and effects thereof will be understood in more detail with reference to specific embodiments below.

According to the mobile robot and its location recognition method according to an embodiment of the present invention, it is possible to quickly recognize the location of the mobile robot using a camera and a laser sensor.

According to the mobile robot and its location recognition method according to an embodiment of the present invention, instead of comparing the current frame and the previous frame as a whole, a part of the previous frame is compared to apply the ICP algorithm, thereby recognizing the location of the mobile robot. processing speed can be improved.

1 is a block diagram showing a mobile robot according to an embodiment of the present invention.
2 is a block diagram showing a location recognition unit of a mobile robot according to an embodiment of the present invention.
3 is a schematic diagram illustrating setting a region of interest of a mobile robot according to an embodiment of the present invention.
4 is a schematic diagram showing edge detection of a mobile robot according to an embodiment of the present invention.
5 is a schematic diagram showing generation of an edge profile of a mobile robot according to an embodiment of the present invention.
6 is a schematic diagram showing extraction of partial feature information of a mobile robot according to an embodiment of the present invention.
7 is a schematic diagram illustrating pose and position recognition of a mobile robot according to an embodiment of the present invention.
8 is a flowchart illustrating a location recognition method of a mobile robot according to an embodiment of the present invention.

Hereinafter, preferred embodiments will be described in detail so that those skilled in the art can easily practice the present invention with reference to the accompanying drawings. However, in describing a preferred embodiment of the present invention in detail, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. In addition, the same reference numerals are used throughout the drawings for parts having similar functions and actions.

In addition, throughout the specification, when a part is said to be 'connected' to another part, this is not only the case where it is 'directly connected', but also the case where it is 'indirectly connected' with another element in between. include In addition, 'including' a certain component means that other components may be further included, rather than excluding other components unless otherwise stated. In addition, terms such as “~unit” and “module” described in the specification refer to a unit that processes at least one function or operation, and may be implemented as hardware or software or a combination of hardware and software.

1 is a block diagram showing a mobile robot according to an embodiment of the present invention.

Referring to FIG. 1 , a mobile robot 100 according to an embodiment of the present invention may include a camera 110, a laser sensor 120, a position recognition unit 130, and a traveling unit 140.

Hereinafter, a mobile robot according to an embodiment of the present invention will be described with reference to FIG. 1 .

The camera 110 may generate image information by photographing the exterior of the mobile robot 100 . One camera 110 may be installed, but depending on embodiments, two or more cameras may be provided to generate a stereo image.

The laser sensor 120 may generate depth information by measuring a distance of an obstacle or the like located around the mobile robot 100 . Depending on the embodiment, the distance to the object may be measured by measuring the time required for the laser beam to be reflected from the object and return to the laser sensor 120 after the laser sensor 120 radiates the laser beam to the object. there is. In general, the laser sensor 120 performs functions such as detecting obstacles located around the mobile robot 100, but here, it recognizes the position, pose, etc. of the mobile robot 100 by combining with image information. can be utilized A laser beam irradiated by the laser sensor 120 may have a dot or line shape, and here, a line-shaped laser beam is referred to as a laser line.

Meanwhile, the laser sensor 120 can measure depth information within a set measurement angle range in front of the mobile robot 100, and within the measurement angle range, measurement points that equally divide the measurement angle range at regular reference angle intervals are provided. may be set. That is, the laser sensor 120 may generate depth information for a plurality of measurement points included in the measurement angle range. For example, the laser sensor 120 may have a range of 100 degrees in front of the mobile robot 100 as a measurement angle range, and 400 measurement points are set to equally divide the measurement angle range into reference angle intervals of 0.25 degrees. There may be. In this case, the laser sensor 120 may measure distances of 400 measurement points within a range of 100 degrees forward and provide the depth information. Here, the measurement angle range and the reference angle interval of the laser sensor 120 may be set in various ways according to the type of the laser sensor 120 and the like.

The position recognition unit 130 may estimate the position and pose of the mobile robot 100 using image information and depth information received from the camera 110 and the laser sensor 120 . In order to set a driving path of the mobile robot 100 or to actually drive the mobile robot 100, it is necessary to grasp the current position and pose of the mobile robot 100. To this end, the position recognition unit 130 may recognize the surrounding environment of the mobile robot 100 using the camera 110 and the laser sensor 120, and the pose and position of the mobile robot 100 from the recognized surrounding environment. Location can be updated (SLAM: Simultaneous Localization and Mapping).

Specifically, as shown in FIG. 2 , the location recognizing unit 130 may include a matching unit 131, a feature information generating unit 132, and a location estimating unit 133.

The matching unit 131 may receive image information and depth information from the camera 110 and the laser sensor 120, respectively, and may calibrate the received image information and depth information. Since the camera 110 and the laser sensor 120 have different installation positions within the mobile robot 100, their respective coordinate systems may be set differently. That is, the same position in the image information and the depth information may be displayed as different coordinates. Therefore, the matching unit 130 may perform coordinate conversion according to the installation positions of the camera 110 and the laser sensor 120, and each image information and depth generated by the camera 110 and the laser sensor 120 Matching coordinates can be set so that the coordinates of the information match.

The feature information generator 132 may set a region of interest (ROI) within the image information, and generate feature information by combining depth information with an edge profile extracted from the region of interest.

First, the feature information generating unit 132 may set a region irradiated with a laser line of the laser sensor 120 as a region of interest among image information. That is, since the feature information generating unit 132 limits the region of interest of the image information to the region irradiated with the laser line, the amount of computation required in a process such as image processing of the image information can be reduced. Here, since the coordinates of the image information and the depth information are matched by the matching unit 110, the laser sensor 120 can easily extract the coordinates of the image information corresponding to the area irradiated with the laser line by using the matching coordinates. there is. After that, as shown in FIG. 3, a region of interest (a) can be set for a range spaced apart by a set interval based on the laser line in the image information.

Meanwhile, the feature information generator 132 may extract an edge profile from the region of interest of the image information, and generate feature information corresponding to the region of interest by combining depth information with the edge profile. there is.

Specifically, the feature information generating unit 132 may apply a Sobel mask to the region of interest to extract edges included in the region of interest, and generate an edge profile using the extracted edges. there is. Depending on the embodiment, a Sobel mask may be formed as shown in FIG. 4(a), and pixel values of pixels included in the region of interest (a) may be differentiated in the x-axis direction using the Sobel mask. . When there is an edge formed in the vertical direction, since pixel values in the x-axis direction change rapidly, a differential value may appear high. Accordingly, an edge image including an edge of a vertical component included in the ROI (a) may be generated by differentiating the ROI in the x-axis direction. Depending on embodiments, it is also possible to generate a binary image by setting a threshold in the region of interest, and then extract an edge image by applying a Sobel mask to the binary image.

Then, after converting the edge image of the region of interest to gray scale, as shown in FIG. ) to create a histogram. In this case, an edge profile may be extracted using the number of bins included in the generated histogram.

When the edge profile is extracted, the feature information generating unit 132 may generate feature information by adding depth information to coordinates corresponding to the measurement point in the edge profile. That is, as shown in FIG. 5(b), feature information can be displayed. Through this, the mobile robot 100 can detect obstacles at high speed from image information obtained while driving. In addition, since feature information can be quickly generated for variable obstacles on the driving path, the surrounding map can be quickly updated based on the feature information.

The position estimation unit 133 may compare feature information at time t and partial feature information at time t-1 using an Iterative Closest Point (ICP) algorithm, and through the comparison result, the pose of the mobile robot 100 ( pose) and position can be estimated. That is, by comparing the feature information around the mobile robot 100 at time t with the feature information around the mobile robot 100 at time t-1, the movement distance and rotation angle of the mobile robot 100 at time t etc. can be grasped, and through this, the pose and position of the mobile robot 100 can be estimated. However, although it is possible to estimate the pose and position using all feature information at the time point t-1, in this case, problems such as an increase in the amount of calculation may occur.

To solve this problem, the location estimation unit 133 may use partial feature information at time t-1 instead of all feature information at time t-1. Here, the partial feature information corresponds to including only feature information of measurement points included in a preset limit angle range among the measurement angle ranges of the laser sensor 120 . That is, as shown in FIG. 6 (a), when the measurement angle range of the laser sensor 120 is 100 degrees, partial feature information including only feature information included in the limit angle range of 50 degrees can be generated, At this time, the partial feature information corresponds to the area shown in FIG. 6(b).

Meanwhile, the location estimator 133 may utilize a point cloud model when estimating the pose and location of the mobile robot 100 . The point cloud model displays a plurality of feature information on a space and can be represented as shown in FIGS. 6 and 7 . The point cloud model can be displayed on a three-dimensional space, but can also be displayed on a two-dimensional space, as shown in FIGS. 6 and 7 .

Specifically, the location estimation unit 133 may extract a target area corresponding to partial feature information at time t-1 from a point cloud model at time t using an iterative closest point (ICP) algorithm. Thereafter, the pose and position of the mobile robot 100 may be estimated by comparing the location of the extracted target region at time t-1 with the location at time t.

로 표현할 수 있으며, t-1 시점의 부분 특징정보에 대응하는 타겟 영역은

으로 표현할 수 있다. 여기서, Q>P, ∀P, Q ∈ [1,...,N]를 만족할 수 있다. Depending on the embodiment, the point cloud model at time t

It can be expressed as, and the target area corresponding to the partial feature information at time t-1 is

can be expressed as Here, Q>P, ∀P, Q ∈ [1,...,N] can be satisfied.

In this case, the measurement angle range is the angle between the first measurement point (i=1) and the N-th measurement point (i=N), and the limit angle range is the P-th measurement point (i=P) to the Q-th measurement point ( It corresponds to the angle between i=Q). For example, when the measurement angle range is 100 degrees in front of the mobile robot 100 and the limit angle range is 50 degrees in front of the mobile robot, N = 400, P = 100, and Q = 300 can be set.

,

일 수 있으며, d⁽ⁱ⁾ _t-1은 t-1시점에서 i번째 측정포인트의 이동거리, Φ⁽ⁱ⁾ _t-1은 t-1 시점에서 i번째 측정포인트의 회전각에 해당한다. A는 각각의 측정포인트 사이의 각도차로, 측정포인트의 개수가 400개이고, 측정각도가 이동로봇(100)의 전방 100도인 경우에는 0.25도로 설정할 수 있다. here,

,

, d ⁽ⁱ⁾ _t-1 corresponds to the movement distance of the ith measurement point at time t-1, and Φ ⁽ⁱ⁾ _t-1 corresponds to the rotation angle of the ith measurement point at time t-1. A is the angle difference between each measurement point, and can be set to 0.25 degrees when the number of measurement points is 400 and the measurement angle is 100 degrees in front of the mobile robot 100.

Therefore, as shown in FIG. 7, the location estimation unit 133 can search the target area at time t-1 in FIG. 7 (a) within the point cloud model at time t in FIG. 7 (b). there is. Here, the ICP algorithm searches for a target area through repeated execution, and since the location estimation unit 133 uses partial feature information, the operation speed of the ICP algorithm can be improved.

In addition, the position estimator 133 reduces the number of iterations of the ICP algorithm by presetting the limiting angle range of the partial feature information according to the maximum movement distance and maximum rotation angle that the mobile robot 100 can move in each frame. can That is, the target area can be designed to be necessarily included in the point cloud model at time t.

에서, E_estimate가 최소값이 되는 이동거리 d 및 회전각 Φ를 추출할 수 있으며, 추출된 이동거리 d 및 회전각 Φ를 이동로봇(100)의 최대이동거리 및 최대회전각으로 설정할 수 있다. Specifically,

In , the movement distance d and rotation angle Φ at which E _estimate is the minimum value can be extracted, and the extracted movement distance d and rotation angle Φ can be set as the maximum movement distance and maximum rotation angle of the mobile robot 100 .

The driving unit 140 may set a driving route according to the pose and position of the mobile robot 100 and drive the mobile robot 100 to move according to the driving route. The driving unit 140 may control a motor, an engine, or steering for driving the mobile robot 100 . Here, the driving unit 140 may directly set a driving path according to the pose and position of the mobile robot 100 and surrounding obstacles for autonomous driving, and according to an embodiment, the mobile robot ( 100) can also be controlled to move.

Specifically, the driving unit 140 may generate a probability-based surrounding map using the surrounding environment recognized by the location recognition unit 130, and may set a driving route to move on the generated surrounding map. At this time, the driving unit 140 may autonomously drive by utilizing a tangent bug algorithm.

8 is a flowchart illustrating a location recognition method of a mobile robot according to an embodiment of the present invention.

Referring to FIG. 8 , the location recognition method of a mobile robot according to an embodiment of the present invention may include a matching step (S10), feature information generation step (S20), and location estimation step (S30).

Hereinafter, a location recognition method of a mobile robot according to an embodiment of the present invention will be described with reference to FIG. 8 .

In the matching step ( S10 ), image information and depth information of time t may be received from the camera and the laser sensor, respectively, and the received image information and depth information may be calibrated. Since the cameras and laser sensors have different installation positions in the mobile robot, their respective coordinate systems may be set differently. Therefore, through the matching step (S10), coordinate conversion can be performed according to the installation positions of the camera and laser sensor, and matching coordinates can be set so that the coordinates of each image information and depth information generated by the camera and laser sensor coincide. can

In the feature information generating step (S20), first, a region of the image information to which the laser line of the laser sensor is irradiated may be set as a region of interest (ROI). Here, since the coordinates of the image information and the depth information are matched by the matching step (S10), the laser sensor can easily extract the coordinates of the image information corresponding to the area irradiated with the laser line by using the matching coordinates. In this case, a range separated by a set interval based on the laser line in the image information can be set as the region of interest.

Thereafter, feature information at time t may be generated by combining depth information with an edge profile extracted from the region of interest. Specifically, edges included in the ROI may be extracted by applying a Sobel mask to the ROI, and an edge profile may be generated using the extracted edges. Depending on embodiments, pixel values of pixels included in the region of interest may be differentiated in the x-axis direction by using the Sobel mask. When there is an edge formed in the vertical direction, since pixel values in the x-axis direction change rapidly, a differential value may appear high. Accordingly, an edge image including an edge of a vertical component included in the ROI may be generated by differentiating the ROI in the x-axis direction.

Then, after converting the edge image of the region of interest to gray scale, neighboring pixels in eight directions can be inspected, and a histogram can be created by assigning bins to directions in which edge components appear strongly. can In this case, an edge profile may be extracted using the number of bins included in the generated histogram.

When the edge profile is extracted, feature information may be generated by adding depth information to coordinates corresponding to measurement points among the edge profiles.

In the location estimation step (S30), feature information at time t and partial feature information at time t-1 may be compared using an iterative closest point (ICP) algorithm. Here, it is possible to estimate the pose and position of the mobile robot through the comparison result. That is, by comparing the feature information around the mobile robot at time t with the feature information around the mobile robot at time t-1, the movement distance and rotation angle of the mobile robot at time t can be grasped. The pose and position of the robot can be estimated. However, although it is possible to estimate the pose and position using all feature information at the time point t-1, in this case, problems such as an increase in the amount of calculation may occur.

To solve this problem, in the location estimation step (S30), partial feature information at time t-1 may be used instead of the entire feature information at time t-1. Here, the partial feature information corresponds to including only feature information of measurement points included in a preset limited angle range among the measurement angle range of the laser sensor.

Meanwhile, in the position estimation step (S30), a point cloud model may be used when estimating the pose and position of the mobile robot. A point cloud model displays a plurality of feature information on a space, and can be displayed on a 3-dimensional space, but can also be displayed on a 2-dimensional space depending on embodiments.

Specifically, in the position estimation step (S30), a target area corresponding to partial feature information at time t-1 may be extracted from a point cloud model at time t using an iterative closest point (ICP) algorithm. Thereafter, the pose and position of the mobile robot may be estimated by comparing the location of the extracted target region at time t-1 with the location at time t.

로 표현할 수 있으며, t-1 시점의 부분 특징정보에 대응하는 타겟 영역은

으로 표현할 수 있다. 여기서, Q>P, ∀P, Q ∈ [1,...,N]를 만족할 수 있다. Depending on the embodiment, the point cloud model at time t

It can be expressed as, and the target area corresponding to the partial feature information at time t-1 is

can be expressed as Here, Q>P, ∀P, Q ∈ [1,...,N] can be satisfied.

In this case, the measurement angle range is the angle between the first measurement point (i=1) and the N-th measurement point (i=N), and the limit angle range is the P-th measurement point (i=P) to the Q-th measurement point ( It corresponds to the angle between i=Q). For example, when the measurement angle range is 100 degrees in front of the mobile robot 100 and the limit angle range is 50 degrees in front of the mobile robot, N = 400, P = 100, and Q = 300 can be set.

,

일 수 있으며, d⁽ⁱ⁾ _t-1은 t-1시점에서 i번째 측정포인트의 이동거리, Φ⁽ⁱ⁾ _t-1은 t-1 시점에서 i번째 측정포인트의 회전각에 해당한다. A는 각각의 측정포인트 사이의 각도차로, 측정포인트의 개수가 400개이고, 측정각도가 이동로봇(100)의 전방 100도인 경우에는 0.25도로 설정할 수 있다. here,

,

, d ⁽ⁱ⁾ _t-1 corresponds to the movement distance of the ith measurement point at time t-1, and Φ ⁽ⁱ⁾ _t-1 corresponds to the rotation angle of the ith measurement point at time t-1. A is the angle difference between each measurement point, and can be set to 0.25 degrees when the number of measurement points is 400 and the measurement angle is 100 degrees in front of the mobile robot 100.

The ICP algorithm searches a target area through iterative execution, and the operation speed of the ICP algorithm can be improved by using partial feature information. In addition, the number of iterations of the ICP algorithm can be reduced by presetting a limit angle range of partial feature information according to the maximum movement distance and maximum rotation angle that the mobile robot can move in each frame. That is, the target area can be designed to be necessarily included in the point cloud model at time t.

에서, E_estimate가 최소값이 되는 이동거리 d 및 회전각 Φ를 추출할 수 있으며, 추출된 이동거리 d 및 회전각 Φ를 이동로봇의 최대이동거리 및 최대회전각으로 설정할 수 있다. Specifically,

In , the movement distance d and rotation angle Φ at which E _estimate is the minimum value can be extracted, and the extracted movement distance d and rotation angle Φ can be set as the maximum movement distance and maximum rotation angle of the mobile robot.

The above-described present invention can be implemented as computer readable code on a medium on which a program is recorded. The computer-readable medium may continuously store programs executable by the computer or temporarily store them for execution or download. In addition, the medium may be various recording means or storage means in the form of a single or combined hardware, but is not limited to a medium directly connected to a certain computer system, and may be distributed on a network. Examples of the medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, and ROM, RAM, flash memory, etc. configured to store program instructions. In addition, examples of other media include recording media or storage media managed by an app store that distributes applications, a site that supplies or distributes various other software, and a server. Accordingly, the above detailed description should not be construed as limiting in all respects and should be considered illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

The present invention is not limited by the foregoing embodiments and accompanying drawings. It will be clear to those skilled in the art that the components according to the present invention can be substituted, modified, and changed without departing from the technical spirit of the present invention.

100: mobile robot 110: camera
120: laser sensor 130: position recognition unit
131: matching unit 132: feature information generating unit
133: position estimation unit 140: driving unit

Claims (14)

In the position recognition method of a mobile robot including a camera and a laser sensor,
a calibrating step of receiving image information and depth information at time t from the camera and laser sensor, respectively, and calibrating the received image information and depth information;
Among the image information, a region irradiated with a laser line of the laser sensor is set as a region of interest (ROI), and the depth information is combined with an edge profile extracted from the region of interest to obtain the time t Characteristic information generation step of generating feature information of; and
A position estimation step of comparing feature information at time t and partial feature information at time t-1 using an Iterative Closest Point (ICP) algorithm, and estimating the pose and position of the mobile robot through the comparison result Location recognition method of a mobile robot comprising a.
The method of claim 1, wherein the matching step
The location recognition method of the mobile robot, characterized in that by performing coordinate conversion according to the installation positions of the camera and the laser sensor, and setting matching coordinates so that the coordinates of the image information and depth information coincide.
The method of claim 2, wherein the feature information generating step
A method for recognizing a location of a mobile robot, characterized in that by using the matching coordinates, extracting an area irradiated with the laser line from the image information and setting a range spaced apart by a set interval based on the laser line as the area of interest. .
The method of claim 1, wherein the feature information generating step
A method for recognizing a position of a mobile robot, characterized in that by applying a Sobel mask to the region of interest, extracting an edge included in the region of interest, and generating the edge profile using the extracted edge. .
The method of claim 1, wherein the depth information is
The location recognition method of a mobile robot, characterized in that it includes a distance measurement value to each measurement point that equally divides the measurement angle range of the laser sensor into reference angle intervals.
The method of claim 5, wherein the feature information generating step
Wherein the feature information is generated by adding the depth information to coordinates corresponding to the measurement points among the edge profiles.
The method of claim 6, wherein the partial feature information
A method for recognizing a location of a mobile robot, characterized in that it includes only feature information of measurement points included in a predetermined limited angle range among the range of measurement angles of the laser sensor.
The method of claim 6, wherein the location estimation step
A method for recognizing a position of a mobile robot, characterized in that for generating a point cloud model for the region of interest using the plurality of feature information.
The method of claim 8, wherein the location estimation step
Extracting a target area corresponding to the partial feature information at time t-1 included in the point cloud model at time t, and comparing the position of the target area at time t-1 with the target area at time t, Position recognition method of a mobile robot, characterized in that for estimating the pose and position of the mobile robot.
10. The method of claim 9, wherein the location estimation step

Generating the feature information of the time point t including N measurement points into the point cloud model using

, Q>P, ∀P, Q ∈ [1,...,N] to generate a target area corresponding to the partial feature information at the time point t-1,
At this time, the measurement angle range is an angle between the first measurement point (i=1) and the N-th measurement point (i=N), and the limit angle range is the P-th measurement point (i=P) to the Q-th measurement point. is the angle between (i=Q),
remind

,

, d ⁽ⁱ⁾ _t-1 is the movement distance of the ith measurement point at time t-1, Φ ⁽ⁱ⁾ _t-1 is the rotation angle of the ith measurement point at time t-1, and the angle between each measurement point A method for recognizing a location of a mobile robot, characterized in that the car is A.
The method of claim 7, wherein the location estimation step
A method for recognizing a position of a mobile robot, characterized in that setting a limiting angle range of the partial feature information according to a maximum movement distance and a maximum rotation angle that the mobile robot can move in each frame.
10. The method of claim 9, wherein the location estimation step

In , the movement distance d and rotation angle Φ at which the E _estimate is the minimum value are extracted, and the extracted movement distance d and rotation angle Φ are set as the maximum movement distance and maximum rotation angle. recognition method.
A computer program stored in a recording medium in combination with hardware to execute the method for recognizing the position of a mobile robot according to any one of claims 1 to 12.
a camera generating image information;
a laser sensor generating depth information;
a matching unit for receiving image information and depth information at time t, and calibrating the received image information and depth information;
Among the image information, the region irradiated with the laser line of the laser sensor is set as a region of interest (ROI), and the depth information is combined with an edge profile extracted from the region of interest to obtain the time point t Characteristic information generating unit for generating feature information of; and
Using an Iterative Closest Point (ICP) algorithm, a position estimation unit that compares feature information at time t and partial feature information at time t-1 and estimates the pose and position of the mobile robot through the comparison result Including mobile robots.

Images