KR20180109284A - Autonomous mobile robot to implement logistics automation systems and method for map building and localization of the robot - Google Patents

Abstract

The present invention relates to a method for performing environment scanning and self-localization of an autonomous mobile robot driving in a warehouse, the method comprising the steps of: preparing an actual grip-based map which shows information on surrounding environment by use of actual distance information measured by the autonomous mobile robot driving in a distribution space; forming a revised map by adding an imaginary wall showing an area limiting the driving of the autonomous mobile robot to the actual map; and setting candidate points of a self-position on the revised map by N numbers in case the autonomous mobile robot actually drives in the distribution space, selecting n candidate points by comparing the distance information calculated in a specific direction on the revised map with the actual distance information measured in the specific direction by a distance sensor at the respective N candidate points, and estimating the self-position based on the n candidate points.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an autonomous mobile robot for automating logistics and an environment recognition method and a magnetic location estimation method for the robot,

The present invention relates to a movement control of an autonomous mobile robot and an autonomous mobile robot, and more particularly, to a method of estimating a position of an autonomous mobile robot based on a grid-based map reflecting an area in which environmental conditions are frequently changed, To a self-running robot.

In order to automate logistics, the logistics site where autonomous mobile robots are operated can be a warehouse where many objects are stored, objects stored in a warehouse are taken out, new objects are loaded and loaded, The shape of the surrounding environment to be changed is very severe.

Therefore, if the autonomous mobile robot uses the map constructed by using the information measured at a specific point in time at the logistics site, the self position can not be identified when the position of the object is changed.

As a method to overcome this problem, a method of attaching an artificial mark in the space of a logistics site and recognizing the mark and allowing the autonomous mobile robot to estimate its position is used. However, there is a problem that a plurality of markers are required to be attached to a plurality of positions, and maintenance must be performed to prevent markings from being damaged.

On the other hand, a recently developed autonomous mobile robot has a laser distance sensor that measures a distance from an object to an object by irradiating the laser at various angles from the center, and uses the measured distance information to travel between objects . When the laser distance sensor is used, a map corresponding to the surrounding environment must be provided in advance. The position of the user is estimated in real time on the basis of the measured distance information, and based on the estimated position, Of the vehicle.

However, even in this case, there is still a problem that the position of the object can not be accurately estimated when the position of the object changes in the surrounding environment. Therefore, there is a need for an improved magnetic localization algorithm that allows accurate estimation of its position within a map, regardless of the position of the movable object.

An object of the present invention is to provide an autonomous mobile robot which can automatically create a grid-based map corresponding to the surrounding environment in a space at a logistics site where items are frequently moved, In this paper, we propose a method for estimating the position of a robot with high accuracy using the grid - based map and the laser range sensor.

The present invention relates to an environmental recognition and a self-localization method of an autonomous mobile robot running in a warehouse or the like. The autonomous mobile robot having a laser distance sensor accumulates distance information obtained through actual traveling, A virtual obstruction wall is set in an area where the position of the object is changed in order to cope with the change of the surrounding environment caused by the fluctuation of the object position, Constructing a map, comparing the measured distance information inputted in real time with the corrected map, and estimating the current self position of the autonomous mobile robot.

Further, when the autonomous mobile robot is located in the area where the virtual wall is set, the autonomous mobile robot adjusts the distance information measured by the autonomous mobile robot to be related to the virtual wall, and estimates its own position by referring to the adjusted distance information.

More particularly, the present invention relates to an environment recognition and self-localization method of an autonomous mobile robot for automating logistics, including: a distance sensor for measuring a distance to an object; and an encoder Creating a grid-based actual map by informing the shape of the surrounding environment using actual measured distance information while the autonomous mobile robot is traveling in a logistics space to be autonomously traveling; Constructing a modified map by adding a virtual wall for displaying an area for limiting the running of the autonomous mobile robot on the actual map; (N is a natural number equal to or greater than 1) of candidate positions in the correction map when the autonomous mobile robot is actually traveling in the distribution space, The method comprising: selecting one or a plurality of candidate points by a method of comparing calculated distance information calculated in a specific direction with actual distance information actually measured in the specific direction by the distance sensor; and based on the selected one or plurality of candidate points And estimating its own position.

The step of estimating the self position may include: determining an initial position at which the autonomous mobile robot starts autonomous travel; Calculating a traveling distance traveled by the autonomous mobile robot using the encoder and generating actual distance information in the specific direction from the self-traveling current position; Combining the calculated travel distance to the initial position and adding noise thereto to set N candidate points on the correction map; Calculating the calculated distance information in the specific direction on the modified map for each of the set N candidate points and calculating the difference between the calculated calculated distance information and the actual distance information for each of the candidate points , Selecting n of the candidate points in order of decreasing calculated difference (n is a natural number 1? N? N); And determining a combination of positions of the selected n candidate points as a new magnetic position of the autonomous mobile robot.

The step of estimating the magnetic position may further comprise: a procedure for determining a new magnetic location after the new magnetic location is determined, the method comprising: determining, for each of the n candidate points selected in order of decreasing the calculated difference, Adding the calculated distance information and the noise to set a total of N new candidate points; calculating the calculated distance information in the modified map from each of the N new candidate points in the specific direction; And the actual distance information in the specific direction at the current position and determines the next new magnetic position of the autonomous mobile robot based on the n candidate points selected based on the calculated difference Quot;

In this case, the virtual wall is configured to connect fixed objects on the actual map, and when the distance to the inside of the area or the virtual wall is measured, (K is a natural number of 1 or more), calculates distance components for the specific directions for each of the k pieces of scan distance information, The shortest one of the calculated distance components can be determined as the actual distance information in the specific direction.

The specific direction may be a front direction, a left direction, and a right direction of the autonomous mobile robot.

According to another embodiment of the present invention, there is provided a robot embodied as an autonomous vehicle in a logistics space in a logistics site, the robot comprising: a distance sensor for measuring a distance to an object with respect to a specific direction; A moving means for moving the robot; An encoder for measuring a distance traveled by the robot; Wherein the control unit controls the operation of the moving unit, generates actual distance information for a specific direction based on the measured value of the distance sensor, calculates a moving distance based on the measured value of the encoder, And a virtual map display unit for displaying map information of the virtual map on the actual map, the virtual map displaying the real map on the actual map, (N is a natural number equal to or greater than 1) in the corrected map, the calculated distance information calculated in each of the set N candidate points in the specific direction, Calculates the difference of the generated actual distance information, selects n of the candidate points in order of decreasing the calculated difference (where n is 1? N? N Be), based on the selected candidate point may include a computer for controlling said moving means to estimate the self location.

In addition, when the robot actually travels a section in which the distance to the inside of the area or the virtual wall is measured, the computer: measures the vicinity of the specific direction toward the virtual wall to generate k pieces of scan distance information (K is a natural number equal to or greater than 1), calculates a distance component for the specific direction for each of the k pieces of scan distance information, and calculates the shortest one of the calculated distance components as actual distance information As shown in FIG.

According to the present invention, the autonomous mobile robot accurately identifies its own position in a logistics environment in which the position of an object in the surrounding environment is frequently changed, and follows the safe route. Based on this, the autonomous mobile robot can process the movement of the object in the logistics environment by itself, and it is possible to support the safe transportation of the logistics.

1 is a schematic view showing the configuration and outline of an autonomous mobile robot according to the present invention.
FIG. 2 shows a schematic step of an environment recognition and self-localization method of an autonomous mobile robot according to the present invention.
Figure 3 shows an example of a grid-based initial real map.
Fig. 4 shows a region A where objects are frequently taken out and brought in.
Fig. 5 shows a correction map to which a virtual wall is added to the initial actual map.
6 shows concrete steps of the magnetic localization method.
7 is a view for explaining the concept of the self-localization method.
Fig. 8 shows a method of calculating the distance information to the virtual wall in the left direction when the autonomous mobile robot passes the region A. Fig.

First, with reference to FIG. 1, a configuration and a form of a robot that autonomously travels a logistics space for automating logistics according to the present invention will be schematically described.

The autonomous mobile robot 100 may largely include a laser distance sensor 110, a traveling means 125, an encoder 120, and a computer 150.

The autonomous mobile robot 100 is for transporting objects by autonomously traveling in a logistics space, and can be configured to be capable of placing objects on the upper part thereof or to be able to pull other transportation vehicles.

The laser distance sensor 110 is a means for identifying an object, an object, or an obstacle in front, back, right, left, and up and down directions of the robot 100 and measuring the distance therefrom. The laser distance sensor 110 measures the distance using a laser, but it is merely an example, and may include various types of distance sensors such as ultrasonic waves, sonic waves, ultraviolet rays, infrared rays, and photographed images.

The general laser distance sensor 110 may be configured to irradiate a laser every 0.5 [deg.] Range in a range of about 270 [deg.] To the left and right in the facing direction, and output distance information indicating the measured distance. However, in the present invention, it is assumed that only the distance information of left (0 °), front (90 °) and right (180 °) in the horizontal direction is used as shown in FIG.

The traveling means 125 may comprise, for example, one or more wheels or caterpillar or legs and a motor or link for rotating or driving it.

The encoder 120 can generate movement distance information by measuring the rotation amount of the motor or the wheel and the operation distance of the leg. Alternatively, the encoder 120 may directly analyze the photographed image of the periphery to calculate the moving distance.

The computer 150 obtains the actual distance information generated by controlling the laser distance sensor 110 and controls the operation of the travel means 125 to autonomously travel the robot 100. The distance traveled from the encoder 120 And performs a function of estimating the current self position of the autonomous mobile robot 100 by processing the environment recognition and self position estimation method according to the present invention to be described later.

1 (b) shows an exemplary embodiment of an autonomous mobile robot for automating logistics according to the present invention.

The autonomous mobile robot 100 may be provided with a pedestal on which an article can be placed. In the vicinity of the pedestal, a conveying mechanism for conveying the objects to or from the shelves may be disposed (not shown).

The laser distance sensor 110 may be disposed at least in front of the autonomous mobile robot 100. [

A computer can be mounted inside the autonomous mobile robot 100. Alternatively, the self-position estimation function and the traveling means control function performed by the computer may be implemented in such a manner that they are processed by an instruction transmitted from a server (not shown) connected via a network.

Hereinafter, the details of the environment recognition and self-localization method of the autonomous mobile robot proposed in the present invention will be described.

2, an autonomous mobile robot 100 according to an embodiment of the present invention includes an autonomous mobile robot 100 and an autonomous mobile robot 100. The autonomous mobile robot 100 includes a laser distance sensor 110, Creating an initial realistic grid-based map related to the surrounding environment of the distribution space S using distance information by the grid; A step of constructing a correction map by correcting and supplementing the grid-based initial measurement map based on the prior information about the space where the movement of the object frequently occurs; The autonomic mobile robot 100 compares the measured distance information measured in real time by the laser distance sensor 110 with the distance information calculated by referring to the correction map to estimate the current magnetic position, (S) in the actual logistics space (S).

The first step of the method for constructing the grid-based map for recognizing the environment of the autonomous mobile robot 100 and estimating the self-position based on the grid-based map is that the autonomous mobile robot 100 is controlled by the manager's manual control or other control system Or the obstacle, the furniture, the column, the object, or the obstacle, by using the distance information actually measured by the laser distance sensor in the course of travel while traveling along an arbitrary mark previously arranged in the space, Based initial real map that displays the grid-based initial real map.

The initial real map is a grid-based map composed of grids having arbitrary scales. The space (S) in which the autonomous mobile robot (100) moves is divided into small grids, and each grating has autonomous traveling There is a probability that there is an object that the robot should avoid (the modified map described below may also be the same grid-based map). As described above, the grid-based map can display the actual peripheral shape simply by two-dimensionally arranged grids and the values of the respective grids (for example, having a value of 1 bit or 4 bits) It can take advantages of the laser distance measuring sensor mounted on the robot, and it is able to quickly reflect the change of the environment on the map. The values of the respective grids of the grid-based map composed of the two-dimensional grids can be continuously updated by accumulating the distance information from the surrounding environment received by the autonomous mobile robot 100 to the respective objects.

In the grid-based map, each lattice is denoted by m, and the lattices are set to be composed of N lattices of m ₁ , m ₂ , ... m _N. Each grid m _i can have a probability value between 0 and 1, and this probability value can be denoted by p (m _i ). The closer the p (m _i ) to the grid, the more likely it is that there is an obstacle free space.

When the position and direction of the autonomous mobile robot 100 are represented by the state variable x and the measured value of the laser distance sensor by z, the state variables and measured values accumulated until time t are represented by x _{1: t} and z _{1 : t} , respectively . Therefore, the probability values of each grid of the grid-based map can be expressed as follows.

In the initial state of t = 1, since there is no information previously input for each lattice of the lattice-based map, an intermediate value of 0.5 is assigned to all lattices.

Thereafter, the probability information of each grid is adjusted by applying the distance information measured while the autonomous mobile robot 100 travels, and as a result, an actual map is generated.

An example of an initial real map completed through the above process is shown in FIG.

Here, the white space is caused by grids having a probability of less than 0.5, which can be regarded as a position where no wall or obstacle exists, and is an area in which the autonomous mobile robot can freely move. On the other hand, the black space is caused by grids having a probability of more than 0.5, which means a position where an obstacle is detected from the laser distance sensor, and becomes an obstacle to the movement of the autonomous mobile robot 100. The part without the actual distance information can be displayed in gray or black while maintaining the probability of 0.5. It is possible to prevent the autonomous mobile robot 100 from traveling even when the probability is 0.5.

The second step of the environmental recognition and self-localization method of the autonomous mobile robot includes an operation of modifying the initial real map generated by the actual measurement of the autonomous mobile robot 100 in consideration of the volatility of the logistics environment.

The logistics site usually has a large number of shelves supported by columns, and objects are stored on shelves. These objects can be taken out or replenished from the shelf from time to time even while the autonomous mobile robot is traveling at another location. When the movement of the object occurs in this manner, the map and the object There will be a difference between the actual distance information after the occurrence of the movement.

The autonomous mobile robot 100 estimates its own position by applying the distance information measured in real time to the stored map. When the environment changes due to the movement of the object, the measured distance information is stored When applied to a map, there is a possibility that it will not be able to pinpoint its position on the map.

Therefore, the virtual blocking wall (or the virtual wall) based on the fixed object, which is not constantly changed even if the movement of the object occurs in the observed obstacles of the initial real map, It can be additionally set on the actual map. The setting of the virtual wall may be performed manually by an administrator of the autonomous mobile robot 100 or may be performed by a computer 150 mounted on the autonomous mobile robot 100 or by a self- As shown in FIG. Here, a map to which a virtual wall is added to an initial actual map is referred to as a correction map.

In the present invention, the autonomous mobile robot is preferably driven by a correction map.

The virtual wall can be set by connecting, for example, a column of a shelf in a logistics environment. Fig. 4 shows an area A including a shelf where objects can be frequently taken out and carried in the initial real map. In this case, a portion indicated by a dot in the area A indicated by a rectangle is the actual column of the shelf, and a virtual wall can be formed by connecting these points or defining an extension line of the point.

Fig. 5 shows that the virtual wall B is set in the area A of Fig. The autonomous mobile robot 100 will run on the modified map and when the virtual wall B is traveling in the set area A, the virtual wall B is recognized as an obstacle, Will only travel through the long aisle of.

On the other hand, in the area A in which the virtual wall B is set, the distance information actually measured by the laser distance sensor 110 and the calculated distance information calculated in the correction map are different from each other, It may become impossible to specify. Therefore, the autonomous mobile robot 100 must estimate its own position in a different manner when moving in the area A. This will be described later.

The third step of the environmental recognition and self-localization method of the autonomous mobile robot is to set the plurality of candidate points for the current magnetic position in the modified grid-based map (i.e., correction map) Compares the calculated distance information calculated from the candidate points with the actually measured distance information of the laser distance sensor 110 measured in real time and estimates the self position based on one or more of the candidate points in real time.

The autonomous mobile robot 100 is based on the Monte Carlo Localization method. This method uses the particle filter (partie filter) method as the main algorithm and includes the following steps.

That is, the autonomous mobile robot 100 initializes (S10) after power is applied as shown in FIG. 6, and then predicts a plurality of candidate points for the self position every control period predict (step S20), update each candidate point (S30), select one or more candidate points to estimate its position and orientation (S40), and continue to the next position And resample (S50) a plurality of candidate points again based on the estimated position to estimate the position. In particular, steps S20 to S50 will be repeated. By this procedure, the robot 100 can estimate its own position on the correction map, determine the moving direction and the moving distance with reference to the correction map, and control the operation of the traveling means 125, S) can be autonomously driven.

6 can be understood with reference to the following description and the conceptual diagram of Fig.

(1) initialize step

The initialization step specifies the current magnetic position as a predefined initial position (for example, a charging place) when the power is applied while the autonomous mobile robot 100 is in the stationary state.

(2) a predict stage

로 표시한다. When the autonomous mobile robot 100 moves by an arbitrary distance in any direction, the moving distance can be calculated by referring to the measured value of the encoder 120 mounted on the wheel of the autonomous mobile robot 100. This calculated travel distance is referred to as odometry information

.

라고 하면, 각 파티클은 다음과 같이 계산된다. 여기서, 노이즈가 추가됨으로써, 오도메트리 정보가 결합된 위치의 주변에 설정된 가상의 파티클들이 자기위치의 후보지점으로 설정된다. The odometry information can be combined with the position of the autonomous mobile robot 100 determined in the initial position or the previous position estimation calculation cycle. Then, N particles are constructed by adding noise by a Gaussian random function to a new position where the odometry information is combined with the previous position. This noise

, Each particle is calculated as follows. Here, by adding noise, virtual particles set around the position where the odometry information is combined are set as candidate points of the self position.

In the case of the autonomous mobile robot 100 according to an embodiment of the present invention, it is assumed that N (for example, 50) particles are used. Each particle may mean a position and an attitude of the autonomous mobile robot 100 which is not yet determined but is close to the actual position.

라 하고, 그 파티클의 가중치를

라 하고, 자율주행 로봇(100)의 격자기반 지도상의 위치좌표를

,

라 하고, 평면상의 자세를

라 하면, 시간 t에서의 가중치가 포함된 파티클들 S_t를 다음과 같이 표시할 수 있다.At this time, the i-th particle at time t

And the weight of the particle

, And the position coordinates on the grid-based map of the autonomous mobile robot 100

,

, And the posture on the plane

, The particles S _t including the weight at time t can be expressed as follows.

In the initial state, the position of the autonomous mobile robot 100 will be a predetermined position, and the weights of the particles generated at the beginning are all the same. That is, the weights of each of N particles, i.e., 50 particles, initially generated in the initial state will be the same.

(3) an update step

The weight of each particle is changed according to the degree of matching with the measured value of the current laser distance sensor 110 among the candidates of the position and attitude of the autonomous mobile robot 100 estimated in the prediction step, , And selects one or a plurality of particles close to the magnetic position in accordance with the order of the weights.

The laser distance sensor 110 mounted on the autonomous mobile robot 100 typically spans a wide angle range of 180 degrees or more in the horizontal direction and approximately 270 degrees in the horizontal direction. In addition, since the conventional laser distance sensor 110 is subdivided into several to several tens of angles at an angle of 1 degree with respect to the entire sensing range, several hundreds to several thousands of distance information are generated even in one measurement for the entire sensing range . In addition, since the laser range sensor 110 can repeat the measurement for the entire sensing range from several tens to several tens of times per second, estimating the self position using all the generated distance information is performed by the computer 150 of the autonomous mobile robot, It becomes a large burden.

For this reason, the autonomous mobile robot 100 according to the embodiment of the present invention is configured such that the robot 100 moves in the leftward direction (for example, 0 degrees), in the front direction (for example, 90 degrees) ) Can be used as the distance information.

Meanwhile, in order to estimate the position reached by the movement of the autonomous mobile robot 100, noise is added to a new position generated by combining odometry information at the previous position, so that 50 particles , The position of each particle is specified on the correction map, and the three-way distance information is calculated on the modified map centered on the specified positions (that is, the calculated distance information). 50 calculation distance information can be calculated.

Then, the three-way distance information currently measured by the autonomous mobile robot 100 is compared with each calculation distance information (for example, each of 50), and according to the degree of the error (or distance difference) Adjust the weight. For example, the smaller the error, the greater the weight can be adjusted.

)과의 거리차이

은 다음과 같이 계산할 수 있다.Let the front distance be l _front , the distance in the right direction be l _right , and the distance in the left direction be l _left among the actual distance values measured by the autonomous mobile robot 100 while driving. Then, set the calculated distance information (front, right, left) calculated on each particle in the correction map as vl _front and vl _right vl _left . In this case, the actual distance information at the current position of the autonomous mobile robot 100 and the distance

)

Can be calculated as follows.

When the distance difference is calculated, the set of particles S _i can be summarized as follows.

는 거리차이

에 의하여 결정된다. 거리차이가 작은 파티클은 가중치가 크고, 거리차이가 큰 파티클의 가중치가 작다. 가중치가 큰 파티클의 위치가 자율주행 로봇(100)의 현재 위치일 가능성이 크다.Next, weights are calculated. weight

Distance difference

. Particles with small distance differences have large weights and particles with large distance differences have small weights. It is highly likely that the position of the particle with a large weight is the current position of the autonomous mobile robot 100.

(4) Pose estimate step

When the weight of each particle is calculated in the update step, the top n particles having the highest weight among the N particles can be selected. n is a value smaller than N and may be about 1/10. The position of each of the n selected particles (for example, five particles) may be averaged, and the average value may be determined as the current position and posture of the autonomous mobile robot 100. The estimation of the current position of the autonomous mobile robot 100 is completed through the posture estimation step.

(5) Resample step

The resampling is for estimating a new position (the next position from the previously estimated magnetic position) when the robot 100 further proceeds at the currently estimated magnetic position, and then the osometry information is stored in the previous position And adding new noise to create new particles. That is, in the posture estimation step, the particles having low weight are deleted, and new particles are generated from the particles that have not been deleted due to high weighting (ie, the particles used for the average calculation).

That is, the remaining n particles are deleted while leaving the previously selected n particles, and a plurality of particles are generated by combining the odometry information and noise using each of the n particles selected, and a total of N new particles . The same weight is set again for the N particles thus generated and used as a new position candidate of the autonomous mobile robot 100. Then, it repeats from the update step.

For example, when the top five particles with the highest weight among the first 50 particles are selected and the current position is determined by the average value, the remaining 45 particles are deleted. Then, the autometry robot 100 acquires the odometry information of the autonomous mobile robot 100, and combines the obtained odometry information with each of the five selected particles to generate five new positions. A total of 50 particles are created by adding noise to each new position to create 10 new particles. The calculated distance information can then be calculated from each particle, and these values can be compared with the measured distance information of the laser distance sensor 110 to adjust the weighting. Finally, we select five particles with high weights and estimate the new self position based on this.

The autonomous mobile robot 100 estimates its own position on the modified map in real time and continuously while repeating the above-described prediction → update → posture estimation → resampling step as described above, and based on the estimated magnetic position, And the traveling distance are set so as to make autonomous travel.

Next, a description will be given of a method capable of accurately estimating the self position by compensating for the difference between the actual distance information and the calculation distance information when the virtual wall B is added to the correction map area A in the present invention do.

The autonomous mobile robot 100 according to an embodiment of the present invention uses three-way distance information of front, right, and left directions. However, when the virtual wall B enters the set area (for example, area A) or one of the three directions extends to the virtual wall B, the distance information in a wider range of additional directions I use more. In particular, the distance information in the additional direction is obtained by dividing the distance measured in the additional direction by an extension of the related direction among the three directions and by reducing the distance from the distance sensor 110 to the foot of the waterline, It is used as a way of considering information.

For example, it can be assumed that the virtual wall B is located on the left side of the autonomous mobile robot 100, as shown in Fig. Since the virtual wall B is set along the column position of the shelf in the correction map, the autonomous mobile robot 100 running in the area A where the virtual wall B exists identifies the column of the shelf, The distance to the left should be recalculated by reference.

If there is no object on the inside of the shelf, the actual distance information to the left direction (0 DEG) can not measure the distance to the virtual wall B or the distance to the pillar of the shelf, Will pass through the area where it was located and measure the distance to an object farther away.

Therefore, when passing near the virtual wall B, the distance information measured in the other angular direction should be referred to. For example, if there is a virtual wall in the left direction, the other angles may be arbitrary intervals such as 0 °, 5 °, 10 °, 15 °, or a certain interval such as 0 °, 2 °, 5 °, You can measure the distance in the direction you have. Then, by using a plurality of pieces of distance information obtained from various directions, the virtual wall B is identified as a set reference shelf column, and the imaginary distance to the identified shelf column is regarded as the measured distance information.

In other words, it is assumed that k pieces of distance information are obtained within an arbitrary angle range from the left 0 ° direction of the autonomous mobile robot 100. At this time, it is assumed that the distance of the i-th distance information is length _i and the angle is α _i . If so, the distance from the k pieces of distance information to the foot of the waterline falling on a straight line in the 0 ° direction can be calculated as follows.

At this time, the distance l _left in the left direction of the autonomous mobile robot 100 may be set to a minimum value among p ₁ , p ₂ , ..., p _k .

Repairs to the feet means that it is the shortest and the shortest l _left, l _left it is the shortest may mean that the detection of the shelf columns on the left. Therefore, in this case, the autonomous mobile robot 100 senses its leftward direction in a wide sense, calculates the distance to the virtual wall B based on the detected column of the shelf, It is a method of estimating.

If a virtual wall exists also in the front or right direction of the autonomous mobile robot, l _front , l _right can be calculated by the same method.

Once the l _front , l _right and l _left are calculated from the laser distance sensor measurement values, the self position of the autonomous mobile robot can be estimated according to the steps as described above.

The embodiments of the present invention described above are merely illustrative of the technical idea of the present invention, and the scope of protection of the present invention should be interpreted according to the claims. It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the essential characteristics thereof, It is to be understood that the invention is not limited thereto.

Claims (7)

A distance sensor for measuring a distance to an object, and an encoder for measuring a traveling distance traveled, the method comprising:
Creating a grid-based actual map by informing the shape of the surrounding environment by using actual measured distance information while the autonomous mobile robot travels in a logistics space to be autonomously traveled;
Constructing a modified map by adding a virtual wall for displaying an area for limiting the running of the autonomous mobile robot on the actual map;
(N is a natural number equal to or greater than 1) of candidate positions in the correction map when the autonomous mobile robot is actually traveling in the distribution space, Calculates one or a plurality of candidate points by comparing calculated distance information calculated in a specific direction with actual distance information measured by the distance sensor in the specific direction, and selects one or more candidate points based on the selected one or more candidate points, And estimating an environment of the autonomous mobile robot for automation of the logistics. The method according to claim 1,
The step of estimating the magnetic position comprises:
Determining an initial position at which the autonomous mobile robot starts autonomous traveling;
Calculating a traveling distance traveled by the autonomous mobile robot using the encoder and generating actual distance information in the specific direction from the self-traveling current position;
Combining the calculated travel distance to the initial position and adding noise thereto to set N candidate points on the correction map;
Calculating the calculated distance information in the specific direction on the modified map for each of the set N candidate points and calculating the difference between the calculated calculated distance information and the actual distance information for each of the candidate points , Selecting n of the candidate points in order of decreasing calculated difference (n is a natural number 1? N? N);
And determining a position where the positions of the selected n candidate points are combined as a new self position of the autonomous mobile robot. 3. The method of claim 2,
As a procedure for determining a new magnetic location after the new magnetic location is determined,
Setting a total of N new candidate points by adding a next movement distance and noise to each of the n candidate points selected in descending order of the calculated difference,
Calculating the calculated distance information in the modified map from each of the N new candidate points in the specific direction and calculating a difference between the calculated calculated distance information and actual distance information in the specific direction at the current position And determining a new self position of the autonomous mobile robot based on n candidate points selected on the basis of the calculated difference, wherein the autonomous mobile robot . The method according to claim 1,
Wherein the virtual wall is configured to connect fixed objects on the actual map,
The autonomous mobile robot may be configured such that, when the distance to the inside of the area or the virtual wall is measured,
(K is a natural number equal to or larger than 1) of k pieces of scan distance information by actually observing the vicinity of the specific direction toward the virtual wall,
Calculating distance components for the specific directions for each of the k pieces of scan distance information,
And determining the shortest one of the calculated distance components as the actual distance information in the specific direction. The method according to claim 1,
Wherein the specific direction is a front direction, a left direction, and a right direction of the autonomous mobile robot. A robot embodied as an autonomous robot in a logistics space in a logistics site, comprising:
A distance sensor for measuring a distance to an object in a specific direction;
A moving means for moving the robot;
An encoder for measuring a distance traveled by the robot;
Wherein the control unit controls the operation of the moving unit, generates actual distance information for a specific direction based on the measured value of the distance sensor, calculates a moving distance based on the measured value of the encoder, And a virtual map display unit for displaying map information of the virtual map on the actual map, the virtual map displaying the real map on the actual map, (N is a natural number equal to or greater than 1) in the corrected map, the calculated distance information calculated in each of the set N candidate points in the specific direction, Calculates the difference of the generated actual distance information, selects n of the candidate points in order of decreasing the calculated difference (where n is 1? N? N Be), based on the selected candidate point by including a computer for controlling said moving means to estimate the self-location, environmental awareness and the self location autonomous robot for the automatic distribution having an estimation function. The method according to claim 6,
When the robot actually travels a section in which the distance to the inside of the area or the virtual wall is measured,
(K is a natural number equal to or larger than 1), and calculates a distance component for the specific direction for each of the k pieces of scan distance information , And determines the shortest one of the calculated distance components as actual distance information in the specific direction. The autonomous traveling robot for automation of logistics having the environment recognition and magnetic position estimation function.

Images