KR20180079215A - Position recognition system and method for moving object based on geomagnetic sensor - Google Patents

Abstract

The present invention relates to a position recognition system and method for locating a moving object based on an indoor magnetic field, more particularly, an earth′s magnetic field, without having to construct other infrastructures. The position recognition system includes a geomagnetic sensor generating geomagnetic information having three-dimensional (3D) directional components in a 3D space by measuring the indoor magnetic field, a content addressable memory (CAM) storing information about measurement positions related to a geomagnetic map previously drawn using the geomagnetic information, the 3D geomagnetic information generated by the geomagnetic sensor, and structure information including information about north to the magnetic field, and a calculator estimating a position of the moving object, based on the geomagnetic information generated by the geomagnetic sensor, the structure information including information obtained from the memory, and information about movement of the moving object, using an extended Kalman filter.

Description

TECHNICAL FIELD [0001] The present invention relates to a position recognition system and method for a moving object based on a geomagnetic sensor,

The present invention relates to a position recognition system and method of a geomagnetic sensor-based moving object.

In the case of a mobile robot operating autonomously and performing a given task, a technique for estimating the position of the robot is needed. In the outdoor environment, the current position information of the robot can be obtained by using data obtained from a GPS (Global Positioning System) sensor. In the case of indoor environment, there are researches on location recognition through various methods such as infrared rays, RFID, and wireless LAN, but there are many parts to be improved in terms of infrastructure construction and accuracy.

Of the indoor location recognition methods using infrared rays, RFID, and wireless LAN, infrared is a relatively simple system, and although the price is low, interference occurs due to sunlight and fluorescent lamp (internal illumination), and LOS (Line Of Sight) There are many limitations because it can not be maintained. RFID has the advantages of low sensor prices, but because of the short communication distance, many readers are needed to improve accuracy. The advantage of using a wireless LAN (WLAN) is that it does not require a separate infrastructure because it can use an already established network over the wireless Internet. However, this method has low location accuracy because the radio waves are lost or attenuated due to the indoor environment such as glass and furniture.

Regarding the study of location recognition method using geomagnetic field, ferromagnetic concrete such as internal reinforced concrete and H-beam and power system distort the geomagnetic field inside the building. This distortion can be difficult to determine the direction of the compass due to geomagnetic distortion, but it can be used as a regional feature for location estimation. Haverinen and Kemppainen proposed a position estimation method using a magnetic field intensity value model. They performed location recognition using the MCL algorithm and showed that the experiment was able to locate only with geomagnetic sensors without other sensors.

In order to overcome the limitation of a single sensor and to grasp the exact position, a method combining a geomagnetic sensor and another sensor has been studied. Another study measured the geomagnetic field using a geomagnetic sensor embedded in a smartphone and compared the results with geomagnetism maps made for location recognition in advance. We used inertial measurement equipment (IMU) embedded in the smartphone to calculate control inputs such as the moving distance and direction of the particle filter. The particle filter used for fusion of geomagnetic sensor and other sensor information has a disadvantage in that the number of particles increases and the calculation amount increases, which burdens the system and slows the processing speed.

To use geomagnetic information for location estimation, a database should be created and managed. Dries Vandermeulen created a fingerprint geomagnetism map in order to test whether the geomagnetic field can be estimated indoors. The fingerprint method is a pattern matching method. If the elements of the pattern are too small, high precision can not be expected.

In order to increase the pattern of the fingerprint geomagnetism map, Chung, et al. Estimated location using the equipment made by arranging four geomagnetic sensors. A map was created by measuring the position at 0.05 m from the measurement point every 120 degrees. However, the fingerprint method is less accurate because it compares the pre-made database with the measured information and patterns and finds possibilities for the location to exist.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide a system and method for recognizing the position of a moving object through a geomagnetic field usable in both indoor and outdoor environments, without needing to construct other infrastructures for location recognition.

A position recognition system for a moving object using an indoor magnetic field according to an embodiment of the present invention includes a geomagnetism sensor for generating the geomagnetism information having a three-dimensional direction component in a three-dimensional space by measuring the indoor magnetic field, Dimensional space, information on each measurement position, the three-dimensional observation information, and structure information having information on the north direction of the magnetic field in association with a previously prepared geomagnetic field map, and stores the structure information in a memory using a CAM (Content Addressable Memory) And an operation unit for estimating the position of the moving object by using the geomagnetism information from the geomagnetic sensor, the structure information obtained by accessing the memory, and the movement information of the moving object, and using the extended Kalman filter.

According to another aspect of the present invention, there is provided a method of recognizing a moving object using an indoor magnetic field, comprising the steps of: generating a geomagnetism information having a three-dimensional directional component in a three-dimensional space by measuring the indoor magnetic field, Storing the information on each measurement position, the three-dimensional observation information, and the structure information having information on the north direction of the magnetic field in association with the previously created geomagnetic field map, using the geomagnetic information, And estimating a position of the moving object by using the structure information obtained by approaching the memory and the movement information of the moving object, wherein the storing step stores the structure information using a CAM (Content Addressable Memory) Wherein the step of estimating the position comprises using an extended Kalman filter, And a step of estimating the position of the moving body.

According to the present invention, it is possible to provide a system and method for recognizing the position of a moving object through a geomagnetic field that can be used in indoor and outdoor environments, without needing another separate infrastructure for recognizing the position of the moving object.

1 is a block diagram illustrating a mobile robot according to an embodiment of the present invention.
Fig. 2 is a view showing the actual body of the mobile robot shown in Fig. 1. Fig.
3 is a schematic diagram for explaining a magnetic field vector.
4 is a diagram showing a geomagnetic field map measured in a room.
5 is a view showing various directions of the geomagnetic field.
FIG. 6 is a diagram showing specificities added to the geomagnetic field map shown in FIG.
FIG. 7 is a diagram showing coordinates of a mobile robot and a map of a geomagnetic field.
8 is a diagram showing a flow chart of the approach to the geomagnetic field map memory.
Fig. 9 is a view for calculating the value of the earth's magnetic field.

The terms and words used in the present specification and claims should not be construed as limited to ordinary or dictionary terms and the inventor may properly define the concept of the term in order to best describe its invention It should be construed as meaning and concept consistent with the technical idea of the present invention.

Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents And variations are possible.

Hereinafter, a method of recognizing the position of a moving object using an indoor magnetic field, which is a unique resource of the earth, will be described in detail with reference to the accompanying drawings so that those skilled in the art can readily carry out the present invention. The mobile robot will be described as an example of a moving body, but is not limited thereto.

FIG. 1 is a block diagram of a mobile robot according to an embodiment of the present invention, FIG. 2 is a view showing a realization of the mobile robot shown in FIG. 1, and FIG. 3 is a schematic diagram for explaining a magnetic field vector.

1, a mobile robot 100 according to an embodiment of the present invention includes an operation unit 110, a geomagnetic sensor 120, a memory 130, and a drive unit 140, So that it can grasp its position.

The geomagnetism map is created by using the three-dimensional element and the location information of the geomagnetic field of the room where the mobile robot 100 is located in advance and stored in the memory 130 in a database. The memory 130 uses a CAM (Content Addressable Memory) method for quick access and efficient management.

The driving unit 140 may be a wheel, a motor, a motor driving unit or the like so that the mobile robot 100 can move,

The operation unit 110 estimates the position of the mobile robot 100 using the movement information of the mobile robot 100 and the geomagnetism information. For this, the operation unit 110 uses an extended Kalman filter with a relatively fast and robust processing speed. The operation unit 110 estimates the position and direction of the mobile robot 100 using the system model, and corrects the position using the geomagnetic sensor 120, which is a global sensor.

The geomagnetic sensor 120 can measure three-dimensional direction components in a three-dimensional space like an acceleration sensor. The three-dimensional direction component is represented by three axes of X, Y, and Z, and each axis has an output orthogonal to the direction of the sensor. The definition of each name of the magnetic field vector shown in FIG. 3 is as follows.

Hereinafter, a method of estimating the position using the geomagnetic field information will be described in more detail. The position estimation method can be divided into a method using the output pattern of each axis of the magnetic field and a method using the intensity of the magnetic field. In the embodiment of the present invention, intensity information is used to design a geomagnetic field model. If only the whole intensity information of the magnetic field is used, it is necessary to add the characteristic because the information of the sensor observation model is small and the accuracy is low.

When estimating the position of a person using the earth's magnetic field, the position and direction of the magnetic field sensor attached or possessed by a person may not be fixed, and the height and direction may be changed. However, since the mobile robot 100 shown in FIG. 2 moves in a state of close contact with the floor, the height of the attached geomagnetic sensor 120 is always constant from the bottom and the directions of the mobile robot 100 and the geomagnetic sensor 120 are constant It can be seen that only the direction changes at the same position.

[Table 1] shows the result of measuring the geomagnetic field by rotating the mobile robot 100 by 30 degrees at the same position with the geomagnetic field measurement table. Even if the direction of the mobile robot 100 changes, the magnitude value of the earth's magnetic field, the magnitude value of the horizontal direction, and the magnitude of the vertical direction do not change.

[Table 1]

As shown in Equation (1), three constant components are set as the state vector of the observation model regardless of the direction of the robot.

[Equation 1]

FIG. 4 is a diagram showing an indoor map on which experiments are performed and a geomagnetic field map measured there. After the coordinate axis of the geomagnetic sensor 120 is aligned with the axis of the global coordinate system, the geomagnetic field is measured and a geomagnetic field map is created. The geomagnetic field map is stored in the memory 130 in the form of an array of the following structures. The structure information at each position has information on the measured position, three observation vector information, and information on the north direction of the magnetic field.

The memory 130 uses a CAM (Content Addressable Memory) method for fast operation. CAM is called content addressable memory, and is the main method used in very fast search programs. And stores and reads data on the basis of the ID information of the geomagnetic field coordinates in the memory 130. If you know the address where information is stored in this way, you can access the information of that address at once. On the other hand, if the address is not known, it is necessary to sequentially access from the beginning, read the contents, and compare with the information to be searched repeatedly. This operation is very inefficient because it consumes a memory cycle. Since the arithmetic unit 110 performs the extended Kalman filter operation, it often accesses the memory 130 of the geomagnetic field map and obtains the observation vector and direction information at the predicted position. Therefore, by using the CAM memory 130, the calculation speed of the entire filter can be increased.

In order to verify the algorithm according to the embodiment of the present invention, a laboratory (9.0 m long x 6.6 m long) was modeled and a geomagnetic field map was created. Fig. 4 shows the modeled laboratory view and the position where the geomagnetic field is measured. As shown in [Table 1], a geomagnetic field map was created by measuring the geomagnetic field at 12 points in a horizontal direction, 10 points in a vertical direction, and 120 points in a total of 0.45m intervals inside the laboratory. As a result of recording the north direction of the geomagnetic field, it can be seen that the measured direction indicates different directions.

[Table 1]

[Table 2]

[Table 2] is a table that analyzes the intensity of the earth's magnetic field in the modeled laboratory environment. The magnitude of the earth's magnetic field is 106.3 μT in the largest region and 8.8 μT in the lowest region. The average value of the earth's magnetic field was 55.6 μT.

When the intensity of the magnetic field between the measured positions was compared, the largest change was found to be 31.7 μT and the smallest change was 0.08 μT. The RMS average for change was 10.4 μT. Assuming that the measured interval is 0.45m, and the global magnetic field changes linearly in the interval, it can be said that 0.023μT per 0.01m varies on average.

5 (a) to 5 (d) are views showing various directions of the geomagnetic field. As shown in Fig. 5 (a), it is seen that the compass needle is constantly directed in the direction of the north pole in the form of a geomagnetic field outdoors. However, as shown in FIG. 5 (b), the magnetic field disturbance occurs in the room, and the direction of the compass does not coincide with the polar direction according to the position. However, it can be pointed in various directions as a whole, but it is locally pointed in a distorted direction constantly.

As shown in Figs. 5 (c) and 5 (d), there is an area where the direction of the magnetic field is constant, and it is difficult to determine the direction by the compass even in the outdoors where the direction is easily determined. An Arctic or Antarctic is a region where the pole of the Earth's magnetic field emits or converges. The direction of the magnetic field is perpendicular to the ground, and the N pole of the compass is directed toward the sky or the earth. Fig. 6 is a map of a geomagnetic field in which the direction of the compass is distorted in the indoor space or the region where the pole of the magnetic field diverges or converges.

)에서 시스템 모델에 의하여 예측된 이동 로봇(100)의 좌표(

)가 도 7에 표시되어 있다. 해당 좌표 정보를 변환하여 지구 자기장 맵에서 예측된 관측 정보를 읽어온다.The previous position of the mobile robot 100 (

) Coordinates of the mobile robot 100 predicted by the system model (

Is shown in Fig. And converts the coordinate information to read the observation information predicted from the geomagnetic field map.

8 is a flow chart for accessing the CAM. The coordinate (global coordinate system) of the mobile robot 100 is converted into the coordinate system ID of the geomagnetic field map by using the function using the following equation (2).

&Quot; (2) "

The converted x-direction and y-direction ID are converted into one value by using (2). The access address of the corresponding ID is returned through the CAM function, and then the information shown in [Table 3] is read.

[Table 3]

만큼 떨어진 위치마다 지구 자기장을 측정하였다. 지도 속의 점(

)은 시스템 모델에 의하여 예측된 이동 로봇(100)의 위치이다. 센서 관측 모델 함수

는 이동 로봇의 예상 위치에서 관측 모델의 예측값을 구하는 비선형 함수이다.FIG. 9 is a diagram showing a value of a geomagnetic field and is a geomagnetic field map prepared for an experiment. Constant distance

Of the Earth's magnetic field. Point in Map (

Is the position of the mobile robot 100 predicted by the system model. Sensor Observation Model Functions

Is a nonlinear function for obtaining the predicted value of the observation model at the expected position of the mobile robot.

&Quot; (3) "

만큼 떨어진 거리에서의

네 점에서의 관측 벡터값을 [수학식 4]와 같이 선형 보간법을 이용하여 구하고 평균값을 이용하여 예측된 이동 로봇(100)의 위치에서의 관측 벡터의 값을 구한다.In order to reduce the error at the boundary of the map,

At a distance of

Observation vector values at four points are obtained using linear interpolation as shown in Equation (4), and the value of the observation vector at the position of the mobile robot 100 predicted using the average value is obtained.

&Quot; (4) "

네 지점의 변화량을 이용하여 구한다. 방향에 대한 변화량은 구할 수가 없기 때문에 각도에 의한 변화량은 0이 된다. 앞에서 설명한 과정을 통해 확장 칼만 필터의 결과로 이동 로봇(100)의 위치를 추정할 수 있게 된다. 하지만 이동 로봇(100)의 방향을 알 수 없기 때문에 이동 로봇(100)의 방향을 새롭게 결정해주어야 한다.The matrix H is a matrix that linearizes the geomagnetic nonlinear model function with respect to the state vector. This matrix represents the variation of the observation vector with respect to the state vector. Each component

It is obtained by using the variation of four points. Since the amount of change with respect to the direction can not be obtained, the amount of change due to the angle becomes zero. The position of the mobile robot 100 can be estimated as a result of the extended Kalman filter through the process described above. However, since the direction of the mobile robot 100 can not be determined, the direction of the mobile robot 100 must be newly determined.

와 행렬 H를 구할 때 사용했던

의 방향 정보를 이용하여 로봇이 측정한 자기장 방향과 비교하여 방향을 결정한다.5 (b), the direction information of the measured sensor 120 when the geomagnetic field map is created using the characteristic pointing in a locally distorted direction and the direction information of the geomagnetic field sensor of the mobile robot 100 120) information can be used to determine the direction at the global location. function

And the matrix H

The direction is determined by comparing the direction of the magnetic field measured by the robot with the direction information of the robot.

In the designed geomagnetic field model, since the amount of change of the magnetic field with respect to the angle can not be obtained, the position of the mobile robot 100 can be estimated, but the direction can not be estimated. In order to determine the direction of the mobile robot 100, another method should be used. For a mobile robot or a person, a compass is normally used to determine the direction. Determines the orientation of the compass relative to the direction of the compass, generally in the direction of the Arctic, at a location throughout the earth. However, the direction of the room is unknown due to the disturbance of the magnetic field.

In the room, there are various pole points due to the distortion of the magnetic field. Even if the accurate magnetic field value is measured in this area, the direction is lost. As a result, prediction of the next position by the system model becomes meaningless, and the position of the mobile robot 100 is missed. There is a need to find a singular point and take measures to avoid losing direction in that area.

Two methods are used to find the pole. The first is to analyze the geomagnetic field map and consider that there is a singular point between two points when the north direction of the magnetic field at a close point suddenly changes. In the singular point, there exists an S-pole type in which the direction of the magnetic force lines converge and an N-pole type in which the magnetic force lines diverge. In such a singularity, it is judged that there exists a singular point near the region because the magnetic field line changes around the pole.

Second, it is judged that there exists a singular point in the vicinity when the intensity of the horizontal component of the measured magnetic field is smaller than a certain magnitude. The elements that determine the north of the Earth's magnetic field are the x-axis and y-axis components of the magnetic field. At the pole, there is no horizontal strength because the direction of the magnetic field is directed toward the sky or the ground. As a result, if the intensity of the magnetic field in the horizontal direction is close to zero, there is a possibility that the pole exists in the vicinity or the pole changes into the region.

In the singularity found in this way, the direction is determined by using the odometry information of the system model without determining the direction using the direction of the magnetic field.

The foregoing detailed description is illustrative of the present invention. It is also to be understood that the foregoing is illustrative and explanatory of preferred embodiments of the invention only, and that the invention may be used in various other combinations, modifications and environments. That is, it is possible to make changes or modifications within the scope of the concept of the invention disclosed in this specification, the disclosure and the equivalents of the disclosure and / or the scope of the art or knowledge of the present invention. The foregoing embodiments are intended to illustrate the best mode contemplated for carrying out the invention and are not intended to limit the scope of the present invention to other modes of operation known in the art for utilizing other inventions such as the present invention, Various changes are possible. Accordingly, the detailed description of the invention is not intended to limit the invention to the disclosed embodiments. It is also to be understood that the appended claims are intended to cover such other embodiments.

100: Position recognition system
110:
120: Geomagnetic sensor
130: memory
140:

Claims (2)

A system for locating a moving object using an indoor magnetic field,
A geomagnetic sensor for measuring the indoor magnetic field to generate geomagnetism information having a three-dimensional directional component in a three-dimensional space,
Dimensional space, and information on each measurement position, the three-dimensional observation information, and information on the north direction of the magnetic field in association with the geomagnetic field map prepared in advance using the geomagnetism information, ) Way of using memory, and
Estimating a position of the moving object by using the geomagnetism information from the geomagnetism sensor, the structure information obtained by accessing the memory, and the moving information of the moving object,
Containing
Location awareness system. A method of recognizing a position of a moving object using an indoor magnetic field,
The geomagnetic sensor measuring the indoor magnetic field to generate geomagnetism information having a three-dimensional directional component in a three-dimensional space,
Storing the structure information having information on each measurement position, the three-dimensional observation information, and information on the north direction of the magnetic field in association with the previously prepared geomagnetic field map using the geomagnetism information, and
Estimating a position of the moving object using the geomagnetism information, the structure information obtained by accessing the memory, and the movement information of the moving object
/ RTI >
Wherein the storing step includes storing the structure information using a CAM (Content Addressable Memory)
Wherein the estimating includes estimating a position of the moving object using an extended Kalman filter
Location recognition method.

Images