WO2011084012A2 - Global position estimation and correction method of mobile robot using magnetic landmarks - Google Patents

Abstract

The present invention relates to a positional error estimation and correction method of a mobile robot, and more specifically, to a global position estimation and correction method of a mobile robot using magnetic landmarks, in which the mobile robot recognizes patterns of a particular landmark and a neighboring landmark and compares the recognized patterns with a stored landmark space map, if the mobile robot with four hall sensors attached on the bottom surface thereof moves on a mobile plate on which multiple kinds of landmarks formed by changing the arrangement of magnets are disposed, thereby easily estimating coordinate values of an absolute position of a landmark at which the mobile robot is currently positioned, and easily correcting positional errors of the mobile robot. The present invention, that is, the global position estimation and correction method of the mobile robot using the magnetic landmarks, is comprised of constituent means, and the global position estimation and correction method of the mobile robot using the magnetic landmarks comprises the steps of: estimating coordinate values of an absolute position of a landmark at which the mobile robot is currently positioned, by recognizing a pattern of a particular landmark at which the mobile robot is positioned and a pattern of a neighboring landmark; estimating positional errors (x_e, y_e, θ_e) of the mobile robot which deviates from the estimated coordinate values of the absolute position of the landmark; and correcting the positional errors by moving the mobile robot as much as the estimated positional errors (x_e, y_e, θ_e), so that the mobile robot may move to the coordinate values of the absolute position of the landmark.

Description

Global Position Estimation and Correction Method of Mobile Robot Using Magnetic Landmarks

The present invention relates to a method for estimating and correcting a position error of a mobile robot, and in particular, a moving plate having a plurality of kinds of landmarks formed by changing the arrangement of magnets of a mobile robot having four Hall sensors attached to the floor. In the case of moving, the mobile robot recognizes a pattern of a specific landmark and a neighboring landmark, and compares it with the stored landmark space map, whereby the absolute position coordinates of the landmark where the mobile robot is currently located. The present invention relates to a global position estimation and correction method of a mobile robot using a magnetic landmark that can easily estimate a value and can easily correct a position error of the mobile robot.

Recently, with increasing interest in robots and technological developments, robots that can be applied to real life are appearing one after another. Such a robot performs a given role based on the movement. The hardware structure for the movement is based on the mobile robot, and since the movement is the main role, the position recognition technology for the autonomous movement is essential.

However, since the robot was developed due to the difficulty in implementing the location recognition technology, the location recognition technology is still being studied.

A typical mobile robot uses an odormetery to recognize a position by driving two wheels respectively. However, the odometry has an error, and the error increases with time due to the accumulation of errors. To solve this problem, many studies have been conducted. For example, the odometry is analyzed by analyzing the hardware characteristics of the robot through experiments. There is a University of Michigan Benchmark test (UMBmark) to remove the error.

In addition, many sensors other than odometry have been fused to improve the location recognition method through probability and statistical approaches. In addition, a lot of researches have been conducted to solve the problem of location recognition problem such as recognizing a specific location by using RFID and recognizing it or analyzing the location through image processing using a vision camera.

However, accurate position error estimation and correction of the mobile robot is still a problem to be studied continuously, it is difficult to accurately determine the absolute position, and development through another approach is required.

Meanwhile, in a mobile space in which a mobile robot moves, a method of estimating an absolute position in which a mobile robot currently exists also needs research and development.

The present invention was devised to solve the above problems of the prior art, a mobile robot having four Hall sensors attached to the bottom surface is a movement in which a plurality of kinds of landmarks formed by changing the arrangement of the magnets are arranged In the case of moving the plate, the mobile robot recognizes a pattern of a specific landmark and neighboring landmarks, and then compares it with the stored landmark space map, whereby the absolute position of the landmark where the mobile robot is currently located. An object of the present invention is to provide a global position estimation and correction method of a mobile robot using a magnetic landmark that can easily estimate coordinate values and easily correct a position error of the mobile robot.

In order to solve the above problems, the constituent means of the global position estimation and correction method of the mobile robot using the magnetic landmark of the present invention is a mobile robot in the global position estimation and correction method of the mobile robot using a hall sensor. Estimating an absolute position coordinate value of a landmark currently located by the mobile robot by recognizing a pattern of a specific landmark located and a neighboring landmark, wherein the deviation from the absolute position coordinate value of the estimated landmark is determined. Estimating a position error (x _e , y _e , θ _e ) of the mobile robot, such that the estimated position error (x _e , y _e ,) can be moved to an absolute position coordinate value of the landmark. and _e ) correcting the position error by moving the mobile robot by θ _e ).

The estimating an absolute position coordinate value of the landmark may include a first process of recognizing a pattern of the specific landmark from a specific landmark where the mobile robot is located and then moving to a neighboring landmark, wherein the movement A second process of recognizing a pattern of a landmark in a neighboring landmark, a third process of comparing the pattern of the recognized neighboring landmark and a pattern of the specific landmark with a landmark space map, and the comparison result And when there is only one landmark group pattern composed of the neighboring landmark pattern and the specific landmark pattern in the landmark space map, the landmark of the landmark on which the mobile robot is currently located on the landmark space map. Extract absolute position coordinates and, if two or more exist, move to another neighboring landmark And a fourth process of performing the second process after the movement.

Here, the landmark space map is configured to correspond to the respective patterns of the landmarks arranged on the moving plate which is the space in which the mobile robot moves, and the patterns of the landmarks are n magnet positions constituting each landmark. It is characterized in that the P pattern formed by changing the.

이고, 여기서, N은 자석의 개수이고, ₁H_N은 N개의 자석을 한 방향으로 놓는 경우의 수를 의미하는 중복조합이고, m은 자석의 배치 방향의 경우의 수를 의미하고, gcd(a, N-a)는 a와 N-a의 최대공약수이고, l│gcd(a, N-a)는 l이 gcd(a, N-a)의 약수임을 의미하며,

는 l 이하의 l과 서로소인 자연수의 개수를 의미하는 오일러 함수임.Where P is

Where N is the number of magnets, ₁ H _N is an overlapping combination that means the number of cases where N magnets are placed in one direction, m is the number of cases in the direction of magnet placement, and gcd (a , Na) is the greatest common divisor of a and Na, and l│gcd (a, Na) means that l is a divisor of gcd (a, Na),

Is an Euler function that means the number of natural numbers that are less than or equal to l less than l.

Herein, the mobile robot is configured to store one-to-one matching of the patterns of the P landmarks and the output voltage when the mobile robot rotates on the P landmarks.

According to the global position estimation and correction method of the mobile robot using the magnetic landmark of the present invention having the above problems and solving means, the mobile robot having four Hall sensors attached to the bottom surface is formed by changing the arrangement of the magnets. When moving a moving plate on which a plurality of kinds of landmarks are arranged, the mobile robot recognizes a pattern of a specific landmark and a neighboring landmark, and then compares it with the stored landmark space map, thereby the mobile. The advantage is that the absolute position coordinates of the landmark where the robot is currently located can be easily estimated.

In addition, there is an advantage that can easily estimate and correct the position error of the mobile robot deviating from the absolute position coordinate value of the estimated landmark.

1 is a perspective view of a mobile robot applied to the present invention.

Figure 2 is a detailed view showing the bottom surface of the mobile robot applied to the present invention.

3 is an understanding diagram for explaining the relationship between the Hall sensor output voltage and the Hall sensor position applied to the present invention.

4 is an exemplary view of a moving plate applied to the present invention.

5 is an exemplary view illustrating a landmark space map in which a plurality of landmarks are formed on a moving plate applied to the present invention and are stored corresponding to the landmarks.

6 is a detailed view showing patterns of landmarks applied to the present invention.

7 is an exemplary view for explaining a procedure of extracting absolute position coordinate values applied to the present invention.

8 is an exemplary view for explaining a landmark coordinate system applied to the present invention.

9 is an understanding diagram for explaining a position error estimation applied to the present invention.

10 is a flowchart illustrating a global position estimation and correction method of a mobile robot using magnetic landmarks applied to the present invention.

11 is an understanding diagram for explaining a position error estimation applied to the present invention.

12 is an understanding diagram for explaining a position error value applied to the present invention.

13 is a flowchart illustrating a position error correction procedure applied to the present invention.

14 is another operation procedure for explaining a position error correction procedure applied to the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the global position estimation and correction method of the mobile robot using the magnetic landmark of the present invention having the above problems, solving means and effects.

Before describing the global position estimation and correction method of the mobile robot using the magnetic landmark of the present invention, the environment and conditions to be achieved in order to achieve the present invention will be described.

Mobile robots are generally two wheels or four wheels, and two-wheeled mobile robots are most widely used. A mobile robot driven by two wheels can dynamically express the movement of the robot using the angular velocity of both wheels. In the present invention, the two wheels shown in FIGS. 1 and 2 are used in a mobile robot of a differential-drive.

As shown in FIG. 2, in the mobile robot 10 having two wheels applied to the present invention, two wheels 1 protrude outside the bottom surface. These two wheels 1 are in contact with the upper surface of the moving plate formed with a plurality of landmarks to be described later to allow the mobile robot to move.

On the other hand, as shown in Figure 2, the bottom surface of the mobile robot, a straight line connecting the two wheels (1) (hereinafter referred to as the "center axis of two wheels") perpendicular to the vertical direction (hereinafter "mobile robot") Two ball casters 2 are mounted on the "moving direction line". When the mobile robot moves by such a ball caster 2, the mobile robot can be prevented from tilting in the front-back direction.

Meanwhile, a plurality of Hall sensors 3 are attached to the bottom surface of the mobile robot 10. 2 illustrates a state in which four Hall sensors 3 are attached. Of the four Hall sensors 3, two are attached on the central axis (straight line connecting the two wheels) of the two wheels, the other two are attached to the mobile robot movement direction line. As a result, the four Hall sensors 3 are arranged in a rectangular shape, and the distance between the respective Hall sensors facing each other (the distance between the two Hall sensors attached on the center axis of the two wheels and the mobile robot). The distance between the two Hall sensors attached by the direction of movement of is equal to.

A plurality of Hall sensors 3 (four Hall sensors in FIG. 2) attached to the bottom surface of the mobile robot 10 may detect a magnetic field of each magnet constituting a landmark as a voltage. Perform the function. The Hall sensor 3 is a magnetic sensor manufactured based on the Hall effect, and can detect magnetic flux flux density as a voltage in a magnetic field. The Hall sensor 3 reacts linearly according to the strength of the magnetic field, and outputs a stable voltage even in a fine magnetic field, thereby increasing accuracy when applied to position recognition.

Here, in order to examine the linearity of the Hall sensor 3 applied to the present invention, the Hall sensor output voltage detected from the Hall sensor located above the magnet and the position of the Hall sensor (center line of the magnet ("c" in FIG. Look at the characteristic graph (graph shown in Figure 3) showing the relationship of the position) of the Hall sensor spaced apart from the indicated line).

In the present invention, since the mobile robot having two wheels moves a moving plate made up of a plurality of landmarks, when the mobile robot is located on each of the landmarks, the hall sensor 3 has a magnet and predetermined air. Assume that a gap (air gap, denoted by "h" in FIG. 3) is set to output a voltage between 0 and 5 V by driving a 5 V voltage.

Then, as shown in Fig. 3, when the Hall sensor is not affected by the magnetic field (the Hall sensor is located on the center line of the magnet), it is about 2.47V, and about 4.7V, S, which is the maximum output voltage at the N pole. It can be seen that the minimum output voltage of about 0.3V appears at the pole.

As shown in Fig. 3, "?" Means the horizontal shortest distance between the Hall sensors at the centerline of the magnet. This horizontal shortest distance " l " means the hall sensor position in the graph shown in FIG. The Hall sensor position (horizontal shortest distance from the centerline of the magnet to the Hall sensor) with respect to the Hall sensor output voltage is very little affected by the width direction of the magnet on the magnet surface. For example, no matter where the Hall sensor located above the magnet is on the center line, the voltage detected by the Hall sensor is about 2.47V. As a result, "l", the distance from the centerline of the magnet to the Hall sensor, may be recognized in the longitudinal direction (indicated by "length direction" in FIG. 3) in which the Hall sensor is located at the centerline of the magnet.

Next, a moving plate providing a space in which the mobile robot can move will be described.

4 shows an example of a moving plate applied to the present invention. The moving plate 20 provides a space in which the mobile robot can move. If the mobile robot 10 performs the work in the industrial site, the moving plate 20 may be regarded as the bottom surface of the working space.

The moving plate 20 is provided with a plurality of magnetic landmarks 21 provided to recognize the position of the mobile robot 10 moving as shown in FIGS. 4 and 5. Each magnetic landmark 21 is composed of n (natural numbers) magnets.

The magnets constituting each magnetic landmark are composed of n pieces (where n is a natural number). That is, the magnetic landmark may be composed of a plurality of magnets, and arranged in various forms. The magnetic landmark consisting of the plurality of magnets is in the form of a polygon (a triangle (when the three magnets)), a square (when the four magnets), a hexagon (when the six magnets), etc.) It is preferred to be configured.

4 and 5 illustrate that the magnetic landmark is composed of four magnets. As shown in Figures 4 and 5, the landmark consisting of the four magnets are arranged symmetrically with the magnets. Each landmark 21 maintains a predetermined interval in the form of a grid in the moving plate 20 and has specific coordinates.

The magnets constituting the respective landmarks 21 are configured to be inserted into the moving plate. However, the moving plate 20 is processed in a flat state after the magnets are inserted so as not to interfere with the movement of the mobile robot 10.

As described above, a plurality of landmarks 21 are disposed on the moving plate 20 at predetermined intervals from each other, and each of the landmarks 21 illustrates n magnets (four magnets in FIGS. 4 and 5). ) Magnets are arranged symmetrically with each other.

The landmarks made of the n magnets are P kinds of landmarks that can be formed by changing the position of the magnets. This will be described with reference to FIG. 6 as follows.

이고, 여기서, N은 자석의 개수이고, ₁H_N은 N개의 자석을 한 방향으로 놓는 경우의 수를 의미하는 중복조합이고, m은 자석의 배치 방향의 경우의 수를 의미하고, gcd(a, N-a)는 a와 N-a의 최대공약수이고, l│gcd(a, N-a)는 l이 gcd(a, N-a)의 약수임을 의미하며,

는 l 이하의 l과 서로소인 자연수의 개수를 의미하는 오일러 함수이다.Where P is

Where N is the number of magnets, ₁ H _N is an overlapping combination that means the number of cases where N magnets are placed in one direction, m is the number of cases in the direction of magnet placement, and gcd (a , Na) is the greatest common divisor of a and Na, and l│gcd (a, Na) means that l is a divisor of gcd (a, Na),

Is an Euler function that means the number of natural numbers less than l and l.

The magnetic landmark may be composed of n magnets to form P patterns. More specifically, in the case of constituting the landmark in the form of a triangle with three magnets, four (P) patterns may be formed, and in the case of constituting the landmark in the form of a fire with four magnets, six (P ), And in the case of constituting a hexagonal landmark with six magnets, 14 (P) patterns can be formed. Of course, when constituting a landmark with more magnets, P patterns can be formed.

The number P of patterns that can be formed using three magnets can be obtained according to the following calculation.

The number P of patterns that can be formed using four magnets can be obtained according to the following calculation.

The number P of patterns that can be formed using six magnets can be obtained according to the following calculation.

In this calculation, since 1HN is the number of cases where N magnets can be placed in one direction, it is always 1, and m is the number of magnets in the arrangement direction, so [N / S], [S, N] Always becomes 2. However, as applied in the embodiment of the present invention, it is limited to the case of using a bar magnet having an N pole and an S pole.

6 illustrates the patterns that can be formed when constructing a landmark with four magnets. Referring to this example, patterns that can be formed will be described.

As shown in Fig. 6, each landmark is composed of four magnets. Therefore, there are a total of six patterns that can be formed by modifying the arrangement of the magnets. Among the magnets constituting each landmark, if the polarities of the magnets are arranged in the counterclockwise direction from the magnet on the right side, six patterns as shown in FIG. 6 can be formed. For example, Pattern 1 has a magnet listing pattern of (NS SN NS SN), Pattern 2 has a magnet listing pattern of (NS NS NS SN), and Pattern 3 has (NS SN SN SN) has a magnet listing pattern, Pattern 4 has a magnet listing pattern of (SN NS NS SN), Pattern 5 has a magnet listing pattern of (NS NS NS NS), and the pattern (Pattern) 6 has a magnet ordering pattern of (SN SN SN SN).

When the mobile robot is located in the P landmark patterns, the output voltage patterns output by the plurality of Hall sensors may also be different according to the respective landmark patterns. When the mobile robot rotates or stops in the P landmark patterns, the output voltage patterns output by the plurality of Hall sensors may also be different according to the respective landmark patterns.

As will be described later, the mobile robot stores one-to-one matching of the output voltage patterns when rotating on the respective landmark patterns and the respective landmark patterns (P patterns). Therefore, when the mobile robot is located at a specific landmark, the mobile robot can easily know which of the P patterns the landmark on which it is located is.

On the other hand, the mobile robot stores a layout map of a plurality of landmarks arranged in a moving plate which is a space (work space) to which the mobile robot moves in advance. That is, the layout of a plurality of landmarks composed of P patterns is stored in advance. This is called a landmark space map.

The landmark space map is as shown on the right side of FIG. As shown in FIG. 5, the landmark space map has a map of patterns of 32 landmarks arranged in a specific area (area divided by dotted lines) of the moving plate. As shown in Fig. 5, the mobile robot in the dotted area is a landmark of pattern 5 (one of six patterns that can be formed when constructing a landmark with four magnets) on the landmark space map. It can be seen that the phase is located.

On the other hand, the mobile robot has spatial absolute coordinate information that stores the absolute position coordinate value of each landmark pattern on the landmark space map.

As shown in Fig. 7, assume that the mobile robot located on a specific landmark of the moving plate is at the position of (1). At this time, when the mobile robot is located at a specific landmark (the position of (1)), it can be seen that the specific landmark is pattern 2 (one of six patterns composed of four magnets). Specifically, if the landmark is stopped or rotated, it can be seen that the specific landmark is pattern 2.

Then, when moving to a neighboring landmark ((position (2) or / and position (3)) and stopping or rotating, the landmark at each position is respectively pattern 6 (landmark pattern of position (2)) and pattern. It can be seen that it is 4 (a landmark pattern at position (3)).

As a result, an area in which the landmark of a particular (position (1)) is pattern 2, the pattern of the landmark adjacent to the right is pattern 6, and the pattern of the landmark adjacent to the front pattern is pattern 4, is the only one on the landmark space map. It can be seen that the region (marked with yellow and blue) of, and as a result, the specific landmark is a pattern 2, it can be seen through the previously stored spatial absolute coordinate information that the absolute position coordinate value is (xi, yi). Details of this will be described later.

Meanwhile, the present invention relates to a method for estimating the global position of a mobile robot and correcting an error. That is, it is assumed that a position error of the mobile robot occurs. Here, the position error of the mobile robot is when the mobile robot is located away from each landmark coordinate origin (indicated by “0” in FIG. 8) where the mobile robot is located, and the center of the mobile robot is the landmark coordinate origin. The degree of deviation from. Here, the deviation degree corresponding to the position error includes not only the deviation degree in the x-axis direction and the y-axis direction, but also a spaced angle between the central axis of the two wheels of the mobile robot and the x-axis or the y-axis.

The center of the mobile robot means a center point of a space formed by a plurality of Hall sensors (when composed of four Hall sensors, an intersection point of each virtual line connecting the Hall sensors facing each other).

Therefore, when there is no position error of the mobile robot, the center of the mobile robot is located on the origin of the landmark coordinate system, and the moving direction line (or the center axis of the two wheels) of the mobile robot is the x axis or the landmark coordinate system. when it coincides with the y-axis. Of course, in the absence of such a position error, since the four Hall sensors of the mobile robot are located at the same distance from the center of the mobile robot and located on each magnet, the Hall sensors are each the same centerline of the corresponding magnet. It will be located on the top of the map (when the landmark is made up of four magnets and four Hall sensors are used).

That is, as shown in FIG. 9, the center of the mobile robot (indicated by “O _R ” in FIG. 9) coincides with the origin O of the landmark coordinate system, and the direction of movement of the mobile robot (in FIG. 9). L _m "is coincident with the x-axis or y-axis (illustrated as coinciding with the y-axis in FIG. 9) and each Hall sensor 3 is each disposed on the same position of the centerline of the corresponding magnet. The case is when there is no position error.

The distance from the center of the mobile robot to each hall sensor and the distance from the origin of the landmark to the center of the center line of each magnet may be slightly different, but the same is preferable. In FIG. 9, since the distance from the center of the mobile robot to each hall sensor and the distance from the origin of the landmark to the center of the center line of each magnet are the same, it is determined that each hall sensor is located on the center of the center line of the corresponding magnet. can see.

Here, each magnet corresponding to each Hall sensor or each Hall sensor corresponding to each magnet means a positional error estimation and correction when each Hall sensor is located on the magnet, so that the relationship overlapping each other ( In other words, it corresponds to a case where a specific hall sensor is overlapped when present on a specific magnet.

On the other hand, the present invention is a case where all Hall sensors are located on each corresponding magnet to use the "actual data range" shown in FIG. That is, if any of the four Hall sensors are located out of the corresponding magnets (when located out of the "real data range" in the graph of Figure 3), there is basically a big problem in the mobile robot control system. It is not treated as an error correction problem. In summary, in the present invention, when the mobile robot in motion moves to a specific landmark as a destination and stops, naturally all hall sensors are positioned on each corresponding magnet without departing from the magnet.

Next, embodiments of the global position estimation and correction method of the mobile robot using the magnetic landmark of the present invention in the above-described environment and conditions will be described in detail.

10 is a flowchart illustrating a global position estimation and correction method of a mobile robot using magnetic landmarks according to an exemplary embodiment of the present invention.

When the mobile robot moving the moving plate arrives or lies at a certain landmark, before moving to the landmark corresponding to the next destination, the mobile robot estimates the global absolute position coordinate value of the landmark where the mobile robot is currently located, and the absolute A position error corresponding to the deviation of the mobile robot from the position coordinate value is estimated, and a procedure for correcting the position error is performed.

First, the mobile robot recognizes a pattern of a specific landmark and a pattern of neighboring landmarks and estimates the absolute position coordinate value of the landmark currently located.

The process of estimating the absolute position coordinate value of the landmark where the mobile robot is currently located is as follows.

When the mobile robot is placed on the moving plate, it is placed on a specific landmark. After the mobile robot recognizes the pattern of the specific landmark, the mobile robot moves to the neighboring landmark. The pattern recognition of the specific landmark may be recognized as a stationary state on the landmark or may be recognized by a rotation operation.

As described above, the mobile robot stores the patterns of the plurality of different landmarks and the voltage output patterns output when the patterns are stationary or rotated in the patterns. Therefore, when rotating at the specific landmark, an output voltage is generated through the Hall sensor, and the mobile robot compares the generated output voltage with the stored output voltage patterns. As a result of the comparison, an output voltage pattern having a pattern similar to the generated output voltage may be found, and a landmark pattern stored corresponding to the found output voltage pattern may be recognized as the pattern of the specific landmark.

For example, in FIG. 7, when the mobile robot is placed at a specific landmark at the position of (1) and rotates at this specific landmark, it can be recognized that the pattern of this specific landmark is pattern 2. FIG.

After recognizing the pattern of a specific landmark as described above, the pattern is moved to a neighboring landmark, and the pattern of the landmark is recognized from the moved neighboring landmark. The method of recognizing the pattern of the neighboring landmark is the same as the process of recognizing the pattern of the specific landmark.

For example, in FIG. 7, the mobile robot moves from the specific landmark to a neighboring landmark at position (2), and then in a stationary state or through rotation, the neighboring land at position (2). It can be recognized that the pattern of the mark is pattern 6.

Then, the pattern of the recognized neighboring landmark and the pattern of the specific landmark are compared with a landmark space map. For example, in FIG. 7, the pattern 2 of the specific landmark and the pattern 6 of the neighboring landmark are compared with the landmark space map stored by the mobile robot.

As a result of the comparison, when only one landmark group pattern composed of the neighboring landmark pattern and the specific landmark pattern exists in the landmark space map, the mobile robot on the landmark space map is currently located. If the location coordinate value of the landmark is extracted and two or more exist, the process of recognizing the pattern of the landmark in the neighboring landmark after moving to another neighboring landmark until only one exists, and the landmark The process of comparing the patterns of the landmark map with the landmark space map is repeated.

The landmark group pattern refers to a pattern group of landmarks recognized by a mobile robot in a stationary state or recognized through rotation. That is, in FIG. 7, the pattern 2 of a specific landmark (landmark at position (1)) and a landmark (landmark at position (2)) adjacent to each other in the positive x-axis direction at this particular landmark Pattern 6 is called a landmark group pattern.

As a result of the comparison, if there is only one landmark group pattern composed of the pattern of the neighboring landmark and the pattern of the specific landmark in the landmark space map, the mobile robot on the landmark space map is currently located. Extract the absolute position coordinates of a landmark.

For example, in FIG. 7, a landmark group pattern including a pattern 2 of a specific landmark and a pattern 6 of a neighboring landmark in a positive x-axis direction are compared with the landmark spatial map illustrated in FIG. 7. As a result, if only one matching landmark group pattern exists in the landmark space map, the absolute position coordinate value of the landmark in which the mobile robot is currently located (in this case, the landmark of position (2) in FIG. 7) is obtained. Extract. The mobile robot stores absolute coordinate values of positions of patterns constituting the landmark space map.

However, as a result of the comparison, if two or more landmark group patterns including the pattern of the neighboring landmark and the pattern of the specific landmark exist in the landmark space map, the neighboring landmarks exist until there is only one. After moving to another landmark, the process of recognizing the pattern of the landmark and comparing the pattern of the recognized landmark with the landmark space map are repeated.

That is, in FIG. 7, there are actually a plurality of landmark group patterns composed of the landmark pattern 2 and the landmark pattern 6 in the positive x-axis direction of the landmark pattern 2. In this case, the mobile robot moves to another neighboring landmark. For example, it moves to the landmark at position (3) in FIG. Then, in a stationary state, the landmark pattern is recognized or rotated to recognize the landmark pattern. In FIG. 7, the pattern of the landmark of the position (3) is recognized as the pattern 4. In FIG.

Then, the mobile robot generates a pattern of the neighboring landmarks with the pattern of the specific landmark (pattern 2), that is, the landmark pattern (pattern 6) at position (2) and the landmark pattern at position (3) in FIG. It is checked whether only one landmark pattern group consisting of (pattern 4) exists in the landmark space map.

As a result of the check, the landmark pattern group consisting of the specific landmark pattern 2, the landmark pattern 6 neighboring in the positive x-axis direction, and the landmark pattern 4 neighboring in the positive y-axis direction, are selected from the land shown in FIG. Only one exists in the mark space map.

Therefore, the mobile robot extracts the absolute position coordinate value of the landmark where the mobile robot is currently located (in this case, the landmark of position (3)). As such, when only one landmark group pattern exists in the landmark space map, the mobile robot may know not only the absolute position coordinate values of the landmark currently located but also the absolute position coordinate values of all the landmarks. This is because the mobile robot stores absolute position coordinate values corresponding to all patterns constituting the landmark space map.

According to the procedure described above, not only the absolute position coordinate values of the landmark where the mobile robot is currently located, but also the absolute position coordinate values of other landmarks, the global absolute position coordinate values can be known.

Meanwhile, after the mobile robot recognizes a landmark pattern at a specific landmark, a procedure of estimating and correcting a position error is performed in a process of moving to a neighboring landmark and moving from a neighboring landmark to another neighboring landmark. Can be performed further.

That is, after recognizing a pattern of a specific landmark, and before moving to a neighboring landmark, a procedure of estimating and correcting a position error of a mobile robot located at the specific landmark is performed or a pattern of a neighboring landmark is recognized. Thereafter, before moving to another neighboring landmark, a procedure of estimating and correcting a position error of the mobile robot located at the neighboring landmark may be performed. The procedure for estimating and correcting the position error of the mobile robot may apply the position error estimation and correction procedure described below.

According to the above procedure, after estimating the absolute position coordinate value of the landmark where the mobile robot is currently located, a procedure of estimating and correcting a position error is performed.

When the mobile robot is located on the landmark where the mobile robot is currently located (the landmark of position (3), exemplifying the situation shown in Fig. 7), the plurality of hall sensors each generate a hall sensor output voltage. That is, each of the plurality of Hall sensors attached to the bottom surface of the mobile robot stationary on a landmark composed of four magnets detects a voltage under the influence of a magnetic field generated by the corresponding magnet.

Then, the position error of the mobile robot deviating from the origin of the landmark coordinate system is estimated by using the voltages detected by the plurality of Hall sensors and the Hall sensor output voltage-Hall sensor positional relationship determined by experiment in advance.

In FIG. 7, a voltage is detected using four Hall sensors. In this case, the Hall sensor output voltage-Hall sensor positional relationship is expressed by Equation (1) below.

수식(1)

Formula (1)

This equation is the relationship between the voltage detected by each Hall sensor and the horizontal shortest distance from the centerline of the magnet to the Hall sensor. Where k _i is a linear proportionality constant voltage (V _i) and the horizontal shortest distance (l _i). And since there are four magnets constituting each landmark and four Hall sensors, i is a natural number in the range of 1-4.

The k _i value is a value determined through experiment in advance. Accordingly, when the value V _i is detected, the horizontal shortest distance _i is determined by the Hall sensor output voltage-hall sensor positional relationship.

Specifically, the Hall sensor output voltage-Hall sensor position relation equation is determined through the graph shown in FIG. 3 (characteristic graph regarding Hall sensor output voltage and Hall sensor position). That is, it can be seen the through the graph, the sensor output voltage (V _i) and the horizontal shortest distance (l _i) shown in Figure 3, this time to determine the proportionality constant k _i between the sensor output voltage and the horizontal minimum distance have.

As described above, after four Hall sensors are affected by the magnetic field of each magnet to which they correspond, the Hall sensor output voltage is generated, and then the Hall sensor output voltage-Hall sensor positional relationship is used. The horizontal shortest distances (l ₁ , l ₂ , l ₃ , l ₄ ) between the centerline of each magnet corresponding to the hall sensor can be obtained.

When the horizontal shortest distances l ₁ , l ₂ , l ₃ , l ₄ are obtained, the position error (x _e , y _e , θ _e ) of the mobile robot deviating from the origin of the landmark coordinate system is estimated using these. can do.

Of the component of the position error, x _e is a land means the degree to which the center of the mobile robot away from the origin of the marked coordinates on the x-axis of landmark coordinate system, and y _e is the landmark coordinate system, the origin of landmark coordinate system y-axis Θ _e means the degree of deviation of the center of the mobile robot, θ _e means the degree of inclination of the two wheel center axis of the mobile robot relative to the x axis or y axis of the landmark coordinate system.

By estimating the error of the mobile robot deviating from the origin of the landmark coordinate system using the horizontal shortest distances, the estimated position error may be moved to the origin of the landmark coordinate system. Compensating for the position error by driving the mobile robot as much as possible.

After estimating and correcting the position of the mobile robot in the landmark where the current mobile robot is located through the above-described steps, after moving to the next destination (landmark), the step of estimating and correcting the position error may be performed again. .

However, when the position error value estimated in the step is small enough to be neglected, rather than performing the above steps, moving to the next destination (landmark), and performing the position estimation and correction step may consume power, It is more preferable when considering working speed, position estimation and limitations of correction.

Therefore, it is preferable to further include a position error checking step between estimating the position error of the mobile robot and correcting the position error.

That is, when the position error value is estimated through the above step, it is determined whether the estimated position error is less than or equal to a predetermined error tolerance reference value. The error tolerance reference value is a value determined through a prior experiment and is a boundary value of the position error that can be ignored. The error tolerance reference value may be relatively small when very precise position recognition is required, and may be relatively large in a position recognition system that may be relatively inaccurate.

It is determined whether the estimated position error (x _e , y _e , θ _e ) is equal to or less than a predetermined error tolerance reference value, and when the estimated position error is less than or equal to the error tolerance reference value, the next destination (land Mark), and the above-described step is performed, and if the estimated position error is not equal to or less than the error tolerance reference value (if exceeded), the position error must be corrected. Perform

Here, the meaning that the estimated position error (x _e , y _e , θ _e ) is equal to or less than the predetermined error tolerance reference value means that all of the components of the estimated position error are equal to or less than the error tolerance reference value for the component. For example, if the error tolerance reference values for each component (x _e , y _e , θ _e ) of the position error are determined to be (5 mm, 5 mm, 5 °), respectively, then each component of the estimated position error is the error tolerance mood. It means the case below the each component of a value (for example, (3mm, 3mm, 4 degrees)).

Estimation of the position error (x _e , y _e , θ _e ) of the above-described mobile robot can be obtained by various equations. In the present invention, an error estimation method of the mobile robot will be described in detail with reference to FIG. 11. For reference, FIG. 11 is an exemplary diagram for describing an error estimation in the case of the magnet arrangement corresponding to the pattern 1 among the six patterns shown in FIG. 6.

As shown in FIG. 11, a mobile robot is located on a landmark, but the center of the mobile robot is stopped from the origin O of the landmark where the mobile robot is currently located. At this time, the center of the mobile robot is located at the coordinate values (x _e , y _e ) of the landmark coordinate system, and the four Hall sensors 3 attached to the bottom surface of the mobile robot are respectively shown in FIG. 11, Located on each corresponding magnet (located on the magnet away from the centerline of the corresponding magnet).

The central axis of the two wheels 1 of the mobile robot is distorted by θ _e from the x axis of the landmark coordinate system. Therefore, the position error of the mobile robot is (x _e , y _e , θ _e ).

In order to estimate the position error (x _e , y _e , θ _e ) of the mobile robot, the voltages detected by the Hall sensors and the Hall sensor output voltage-Hall sensor positional relation of the formula (1) Use to find the horizontal shortest distance from the centerline of each magnet to the Hall sensor.

In FIG. 11, the voltage detected by the Hall sensor located on the magnet located on the x axis is V ₁ and the magnetic field of the magnet located on the positive y axis is determined by the magnetic field of the magnet located on the positive x axis. Therefore, the voltage detected by the Hall sensor located on the magnet located on this y axis is V _2, and the Hall sensor located on the magnet located on this x axis by the magnetic field of the magnet located on the negative x axis. The voltage detected by V is V _3, and the voltage detected by the Hall sensor located on the magnet located on this y axis is V _{4 due} to the magnetic field of the magnet located on the negative y axis.

Then, the shortest horizontal distance between the center line of each magnet and the hall sensor corresponding to each magnet can be obtained by the Hall sensor output voltage-Hall sensor positional relational expression of Equation (1). At this time, the obtained horizontal shortest distances become l ₁ , l ₂ , l ₃ , and l ₄ , respectively.

Then, using the l ₁ , l ₂ , l ₃ , l ₄ , the position error (x _e , y _e , θ _e ) is estimated through the following equations (2) to (4).

수식(2)

Formula (2)

수식(3)

Formula (3)

수식(4)

Formula (4)

In the above formula, l ₂ and l _{4, which} are used to find x _e, which is an error of the x axis, are the horizontal shortest distances between the y axis of the landmark coordinate system and the hall sensors respectively corresponding to two magnets located on the y axis. to be. That is, l ₂ is the horizontal shortest distance between the positive y axis and the Hall sensor corresponding to the magnet lying on this y axis, and l _{4 corresponds} to the negative y axis and the magnet lying on this y axis. Is the shortest horizontal distance between Hall sensors.

Further, l ₁ and l ₃ used to find y _e, which is an error of the y axis, are the horizontal shortest distances between the x axis of the landmark coordinate system and the hall sensors respectively corresponding to two magnets located on the x axis. That is, l ₁ is the horizontal shortest distance between the positive x axis and the Hall sensor corresponding to the magnet lying on the x axis, and l _{3 corresponds} to the negative x axis and the magnet lying on the x axis. Is the shortest horizontal distance between Hall sensors.

In addition, d used to obtain an error θ _e corresponding to an angle at which the center line of the two wheels 1 of the mobile robot is spaced based on the x-axis or the y-axis is opposite to each other, as shown in FIG. Distance between the Hall sensors (in the direction). And, l ₁ , l ₂ , l ₃ , l ₄ are the horizontal shortest distances used when estimating x _e and y _e .

As described above, θ _e is an angle between the central axis (denoted as “w” in FIG. 12) of the two wheels of the mobile robot and the x-axis or y-axis of the landmark coordinate system, as shown in FIG. 12. An angle between the central axis w of the two wheels 1 of the mobile robot and the x axis of the landmark coordinate system, and between the central axis w of the two wheels 1 of the mobile robot and the y axis of the landmark coordinate system. Of the angles, the absolute value corresponds to the smaller angle. Accordingly, in the case of FIG. 12, the angle between the central axis w of the two wheels 1 of the mobile robot and the x axis of the landmark coordinate system is selected as θ _e .

In the above formulas (2) to (4), k ₁ , k ₂ , k ₃ , and k ₄ have a value of 1 or -1 depending on the moving direction of the mobile robot in the landmark. Accordingly, the values of k ₁ , k ₂ , k ₃ , and k _{4 have} a value of 1 or -1 depending on the pattern of each landmark and the direction in which the mobile robot moves in each landmark. This is described below.

First, assume that the output voltage relationship according to the position of the Hall sensor located on the magnet follows the graph shown in FIG. That is, it is assumed that the maximum output voltage value is sensed at the N pole of the magnet and the minimum output voltage value is sensed at the S pole of the magnet.

First, among the landmark patterns shown in FIG. 6, the equation for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 1 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through Equations (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. First, due to the nature of the magnet arrangement, it is divided into two cases. First, when the moving direction of the mobile robot moves from the pattern 1 in the x-axis direction (left and right direction in FIG. 11) among the patterns shown in FIG. 6, k ₁ = 1, k ₂ = 1, k ₃ = 1, has a value of k ₄ = 1, and the second is k ₁ = 1, k ₂ when the moving direction of the mobile robot moves from the pattern 1 to the y-axis direction (forward and backward directions in FIG. 11) among the patterns shown in FIG. 6. = 1, k ₃ = 1, k ₄ = 1.

Next, among the landmark patterns shown in FIG. 6, the equation for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 2 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through Equations (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. First, due to the nature of the magnet arrangement, it is divided into four cases. First, when the moving direction of the mobile robot moves in the positive direction of the x-axis (the right direction in FIG. 11) in the pattern 2 among the patterns shown in FIG. 6, k ₁ = -1, k ₂ = 1, k ₃ = 1, k ₄ = 1, the second is k ₁ = 1, k ₂ = 1, k ₃ = -1, k when moving in the positive y-axis direction (upward in FIG. 11) in pattern 2 ₄ = 1, and the third is k ₁ = 1, k ₂ = 1, k ₃ = 1, k ₄ =-1 in pattern 2 when moving in the negative direction of the x axis (left direction in FIG. 11). The fourth has a value of k ₁ = -1, k ₂ = -1, k ₃ = 1, and k ₄ = 1 when moving in the negative direction of the y-axis in the pattern 2 (downward in FIG. 11). .

Next, the equations for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 3 among the landmark patterns shown in FIG. 6 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through the above formulas (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. First, due to the nature of the magnet arrangement, it is divided into four cases. First, when the moving direction of the mobile robot moves in the positive direction of the x-axis (the right direction in FIG. 11) in the pattern 3 among the patterns shown in FIG. 6, k ₁ = 1, k ₂ = -1, k ₃ = 1, k ₄ = 1, and the second is k ₁ = 1, k ₂ = 1, k ₃ = -1, when moving in the positive y-axis direction (upward in FIG. 11) in pattern 3 With a value of k ₄ = -1, the third is k ₁ = -1, k ₂ = 1, k ₃ = 1, k ₄ = in the case of moving in the negative direction of the x-axis (left direction in FIG. 11) in pattern 3 Has a value of -1, and the fourth is k ₁ = 1, k ₂ = -1, k ₃ = 1, k ₄ = -1 when moving in the negative direction of the y-axis in the pattern 3 (downward in FIG. 11). Has a value.

Next, among the landmark patterns illustrated in FIG. 6, the equation for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 4 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through Equations (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. First, due to the nature of the magnet arrangement, it is divided into four cases. First, when the moving direction of the mobile robot moves in the positive direction of the x-axis (the right direction in FIG. 11) in the pattern 4 of the patterns shown in FIG. 6, k ₁ = -1, k ₂ = 1, k ₃ = 1, k ₄ = 1, the second is k ₁ = 1, k ₂ = -1, k ₃ = -1 when moving in the positive y-axis direction (up in FIG. 11) in pattern 4 , k ₄ = -1, and the third is k ₁ = -1, k ₂ = 1, k ₃ = -1, k when moving in the negative direction of the x-axis in the pattern 4 (left direction in Fig. 11) Has a value of ₄ =-1, and the fourth is k ₁ = -1, k ₂ = -1, k ₃ = 1, k ₄ = when moving in the negative direction of the y-axis in the pattern 4 (downward in Figure 11) It has a value of -1.

Next, among the landmark patterns illustrated in FIG. 6, the equation for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 5 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through Equations (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. In this case, only the values of k ₁ , k ₂ , k ₃ , and k ₄ in one case exist regardless of the direction in which the mobile robot moves due to the characteristics of the magnet arrangement. That is, k ₁ = -1, k ₂ = -1, k ₃ = -1, and k ₄ = 1.

Next, the equations for obtaining the position error (x _e , y _e , θ _e ) in the case of the magnet arrangement of the pattern 6 among the landmark patterns shown in FIG. 6 will be described. Even in this case, the position error (x _e , y _e , θ _e ) is basically calculated through Equations (2) to (4).

However, the values of k ₁ , k ₂ , k ₃ and k ₄ vary depending on the direction of movement of the mobile robot. Even in this case, only the values of k ₁ , k ₂ , k ₃ , and k ₄ in one case exist regardless of the moving direction of the mobile robot due to the characteristics of the magnet arrangement. That is, k ₁ = -1, k ₂ = -1, k ₃ = -1, and k ₄ = -1.

As described above, the position is corrected using the estimated position error (correction of moving the center of the mobile robot to the origin of the landmark coordinate system and coinciding the moving direction line of the mobile robot with the x or y axis of the landmark coordinate system). There are two ways to do this.

A first position error correction method will be described with reference to FIG. 13.

First, suppose that the mobile robot estimates the position error in a state where the mobile robot is stationary on the landmark, as shown in Fig. 13A. Then, position error correction is performed using the estimated position error.

In order to correct the position error, as shown in FIG. 13B, the moving direction line of the mobile robot passes through the origin of the landmark coordinate system using the estimated position error values. Rotate the mobile robot. The rotation direction at this time rotates in the direction in which a rotation angle is small. Accordingly, FIG. 13B is a state in which the counterclockwise direction is rotated in FIG. 13A. The rotation angle at this time may be calculated according to various equations using the estimated position error values.

Next, using the estimated position error values, as shown in (c) of FIG. 13, the mobile robot is moved to the origin of the landmark coordinate system. That is, the center of the mobile robot is moved to the origin of the landmark coordinate system. The distance from the origin of the landmark coordinate system to the coordinate point (x _e , y _e ) may be calculated by various equations using the estimated position error values.

Then, using the estimated position error values, as shown in (d) of FIG. 13, the mobile robot so that the movement direction line of the mobile robot coincides with the x axis or the y axis of the landmark coordinate system. Rotate The rotation at this time is to rotate so that the movement direction line of the mobile robot coincides (facing) to the next destination (landmark) movement direction of the mobile robot (indicated by M _D in Fig. 13A). The rotation direction in this case also rotates in a direction having a small rotation angle. Therefore, FIG. 13D is a state in which FIG. 13C is rotated clockwise.

The position error correction step shown in FIG. 13 is a step of correcting the position error on the x-axis and the y-axis in one movement. In other words, in order to move from the coordinate point (x _e , y _e ) of the landmark coordinate system to the origin of the landmark coordinate system, the position error is caused by moving directly to the origin of the landmark coordinate system without performing x-axis movement and y-axis movement separately. Step to calibrate.

Next, a second position error correction method will be described with reference to FIG. 14.

First, as shown in Fig. 14A, assume that the position error is estimated in the state where the mobile robot is stationary on the landmark. Then, position error correction is performed using the estimated position error.

Of the estimated position error value (x _e, y _e, θ _e), using the θ _e, of the movement direction line of the mobile robot, the landmark coordinate system, as shown in Figure 14 (b) Rotate the mobile robot so that it is horizontal to the x axis.

The rotation angle can be obtained using (90 ° -θ _e ). For example, if θ _e is 30 °, the rotation angle will be 60 °. Such a result is that in order to be changed from the state of FIG. 14A to the state of FIG. 14B, it may be rotated clockwise or counterclockwise. (b) It means to go to the state. (B) of FIG. 14 is a state which rotated clockwise in the state (a) of FIG.

When the state shown in FIG. 14B is reached, correction in the x-axis direction is performed. That is, the position error is corrected in the x-axis direction by moving the mobile robot by x _{e in} the x-axis direction. Then, the state shown in FIG. 14C is obtained. Here, the movement in the x-axis direction is a movement for correcting the x-axis of the landmark coordinate system, and as shown in FIG. 14 (c), it means a movement such that the center axes of two wheels of the mobile robot coincide with the y-axis. do.

Since the x-axis correction was performed, the following performs the y-axis correction. To this end, the mobile robot is rotated 90 degrees so that the direction of movement of the mobile robot coincides with the y axis. Then, the state shown in FIG. 14 (d) is obtained. Then, the mobile robot is moved by y _{e in} the y axis direction to correct the position in the y axis direction. Then, the state shown in FIG. 14E is obtained. In this case, the movement in the y-axis direction is a movement for correcting the y-axis of the landmark coordinate system, and as shown in FIG. 14E, a movement in which the center axes of two wheels of the mobile robot coincide with the x-axis. do.

Through this process, the positional error can be corrected by moving the center of the mobile robot to the origin of the landmark coordinate system. In addition, when the position error correction process is performed, since the direction of movement of the mobile robot coincides with the x-axis or the y-axis, the center axes of the two wheels of the mobile robot also coincide with the x-axis or the y-axis.

The above-described procedure is a process of performing position error correction by first performing x-axis correction and then performing y-axis correction. Of course, the position error correction may be performed by performing the y-axis correction first and then performing the x-axis correction. This is briefly described as follows.

That is, among the estimated position error values (x _e , y _e , θ _e ), the mobile robot is positioned such that the direction of movement of the mobile robot is horizontal to the y axis of the landmark coordinate system using θ _e . Rotate Then, in FIG. 14A, the mobile robot is rotated counterclockwise by θ _e . Even in this case, the mobile robot is rotated in a counterclockwise direction so that the direction of movement of the mobile robot is horizontal to the y axis of the landmark coordinate system by rotating at the minimum angle.

Then, correction in the y-axis direction is performed. That is, in the above state, the mobile robot is moved by y _{e in} the y axis direction to correct a position error in the y axis direction. Here, the movement in the y-axis direction is a movement for correcting the y-axis of the landmark coordinate system, and means a movement such that the center axes of the two wheels of the mobile robot coincide with the x-axis.

Since the y-axis correction has been performed, the following performs the x-axis correction. To this end, the mobile robot is rotated 90 ° so that the moving direction line of the mobile robot coincides with the x-axis. Then, the mobile robot is moved by x _{e in} the x axis direction to correct the position in the x axis direction. Here, the movement in the x-axis direction is a movement for correcting the x-axis of the landmark coordinate system, and means a movement such that the center axes of the two wheels of the mobile robot coincide with the y-axis.

Through this process, the position error can be corrected by moving the center of the mobile robot to the origin of the landmark. In addition, when the position error correction process is performed, since the direction of movement of the mobile robot coincides with the x-axis or the y-axis, the center axes of the two wheels of the mobile robot also coincide with the x-axis or the y-axis.

In the case where the x-axis position error correction is performed after the y-axis position error correction is performed as described above, the final state of the mobile robot is a state in which the movement direction line of the mobile robot coincides with the x-axis while the position error correction is completed. Is in.

As described above, the position error correction may be performed in the y-axis correction order after the x-axis correction, or may be performed in the x-axis correction order after the y-axis correction. However, during the position error correction process, by minimizing the total rotation angle of the mobile mobile robot, it is necessary to reduce the power consumption and to improve the working speed or the moving speed of the mobile robot.

Therefore, it is preferable to determine the position error correction order in consideration of the next destination (landmark) movement direction (denoted as M _D in FIG. 14A) in the current landmark which performs the position error correction.

If the next destination movement direction M _D is the x axis direction, the position error correction in the y axis direction is first performed, and then the position error correction in the x axis direction is performed, and the next destination movement direction M _D is y. In the axial direction, the position error correction in the x-axis direction is first performed, and then the position error correction in the y-axis direction is performed.

As such, by determining the position error correction order in consideration of the next destination movement direction, the total rotation angle of the mobile robot in the position error correction process may be minimized. For example, as shown in (a) of FIG. 14, if the next destination movement direction M _D is in the y-axis direction, after performing position error correction in the x-axis direction first, according to the procedure shown in FIG. 14. It is preferable to perform position error correction in the y-axis direction to perform final position error correction.

If the correction is performed according to the procedure shown in FIG. 14 (y-axis correction after x-axis correction), assuming that θ _e is 30 °, the mobile robot rotates by a total of 150 ° in the position error correction process, thereby correcting the position error. To complete and move to the next destination. That is, the rotation angle for the x-axis correction is 60 ° (rotation angle for moving from (a) to FIG. 12 (b) in FIG. 14) and the rotation angle for the y-axis correction is 90 ° (FIG. 14). (C) to (d) in FIG. 14), and when the position error correction is completed, since it is necessary to move to the next destination without additional rotation, the total rotation angle becomes 150 °.

However, on the contrary, after the position error correction is first performed in the y-axis direction, the position error correction is completed by the x-axis direction, and the final position error correction is completed, the mobile robot rotates by 210 ° in order to move to the next destination. Done. That is, the angle to rotate to perform the first y-axis correction is 30 °, the angle to rotate to perform the next x-axis correction is 90 °, in order to face the direction of movement of the mobile robot in the next destination movement direction Since the angle of rotation is 90 °, the total angle of rotation of the mobile robot is 210 °.

As such, the position error correction of the mobile robot which repeats the position error correction and the movement to the next destination (landmark) has the advantage of minimizing the power consumption.

In particular, when a mobile robot having four Hall sensors attached to the floor moves a moving plate on which a plurality of kinds of landmarks formed by changing the arrangement of magnets are moved, the mobile robot is located at a specific landmark and neighbors. By recognizing a pattern of landmarks that exist, and comparing it with the stored landmark space map, the absolute position coordinate value of the landmark where the mobile robot is currently located can be easily estimated, and the position error of the mobile robot can be estimated. A global position estimation and correction method of a mobile robot using a magnetic landmark that can be easily corrected is provided.

Claims (4)

In the global position estimation and correction method of the mobile robot using the magnetic landmark, Recognizing a pattern of a specific landmark where the mobile robot is located and a pattern of neighboring landmarks to estimate an absolute position coordinate value of the landmark where the mobile robot is currently located; Estimating a position error (x _e , y _e , θ _e ) of the mobile robot that deviates from the absolute position coordinate value of the estimated landmark; And correcting the position error by moving the mobile robot by the estimated position error (x _e , y _e , θ _e ) so that the mobile robot can be moved to the absolute position coordinate value of the landmark. Global position estimation and correction method of a mobile robot using a magnetic landmark, characterized in that. The method of claim 1, wherein estimating an absolute position coordinate value of the landmark comprises: A first process of recognizing a pattern of the specific landmark in a specific landmark where the mobile robot is located, and then moving to a neighboring landmark, A second process of recognizing a pattern of a landmark in the moved neighboring landmark, A third step of comparing the recognized neighboring landmark pattern and the specific landmark pattern with a landmark space map; As a result of the comparison, when only one landmark group pattern composed of the neighboring landmark pattern and the specific landmark pattern exists in the landmark space map, the mobile robot on the landmark space map is currently located. Extracting the absolute position coordinate value of the landmark, and if there are two or more, using a magnetic landmark, characterized in that it comprises a fourth process of moving to another neighboring landmark and performing the second process Global position estimation and correction method for mobile robots. The method according to claim 2, The landmark space map, Corresponding to the patterns of the landmarks disposed on the moving plate, which is the space where the mobile robot moves, the patterns of the landmarks are P patterns formed by changing the n magnet positions constituting each landmark. Global position estimation and correction method of a mobile robot using a magnetic landmark, characterized in that. Where P is

Where N is the number of magnets, ₁ H _N is an overlapping combination that means the number of cases where N magnets are placed in one direction, m is the number of cases in the direction of magnet placement, and gcd (a , Na) is the greatest common divisor of a and Na, and l│gcd (a, Na) means that l is a divisor of gcd (a, Na),

Is an Euler function that means the number of natural numbers that are less than or equal to l less than l. The method according to claim 3, The mobile robot stores the patterns of the P landmarks and the output voltages when the mobile robot rotates on the P landmarks in a one-to-one correspondence and store the entire area of the mobile robot using magnetic landmarks. Position estimation and correction method.

Images