KR20180087827A - SLAM method and apparatus robust to wireless environment change - Google Patents

Abstract

The present invention relates to a simultaneous localization and mapping (SLAM) method robust to a wireless environment change and an apparatus thereof. The SLAM method comprises the steps of: estimating a relative position of a mobile node based on motion detection of the mobile node; correcting the relative position of the mobile node based on comparison between a variation pattern of at least one signal strength received over a plurality of time points and signal intensity distribution in a region where the mobile node is positioned; and precisely estimating a position of the mobile node and mapping a map in which highly accurate path information is recorded throughout the region even if change in a wireless environment such as signal interference between communication channels, expansion in access points, and failure or obstacle occur or the wireless environment is poor such as lack of the number of the access points by expressing a path of the region using the corrected relative position.

Description

TECHNICAL FIELD [0001] The present invention relates to a SLAM method and apparatus robust to a wireless environment change,

The present invention relates to a SLAM method and apparatus for estimating a location of a mobile node in an unknown environment and generating a map of an unknown environment.

 Recently, attention has been focused on SLAM (Simultaneous Localization And Mapping) technology in mobile robots such as drone. SLAM is a key technology for autonomous navigation as it is a technology to create a map of unknown environment while recognizing its own position without external help by only the sensor attached to the robot while moving the robot in unknown environment. The conventional SLAM estimates the position of the mobile robot based on several physical landmarks. To identify such landmarks, sensors such as LiDAR, cameras, ultrasonic sensors and the like are required.

Lida has a very high resolution but is expensive and difficult to be applied to small and light devices such as smart phones due to limitations in miniaturization. It is difficult for a camera to be applied to a smart phone or the like having a low processing capability of image data as its output is image data. Ultrasonic sensors can be miniaturized, but resolution is very low and there is a limit to creating high accuracy maps. As a result, SLAM is attracting attention for special purposes such as mobile robots that are put into disaster areas, and is not attracting attention for map generation for positioning of smart phones or vehicle navigation systems in addition to mobile robots.

GPS (Global Positioning System) -based maps and wireless positioning maps are representative examples of maps that have been commercialized or researched for positioning of smart phones or vehicle navigation systems. Such a conventional map has a problem that it takes a lot of time and expense to make a map because a map is created by a person collecting topographic information or signal information while traveling around the area where the positioning service is to be provided. Especially, GPS is impossible to locate in the indoor space where the radio wave transmitted from the satellite can not reach, and there is a problem that the accuracy of the positioning in the city is seriously deteriorated due to the blocking of the radio wave by the high-rise building and the reflection.

In recent years, automobile manufacturers around the world, global companies such as Google and Intel have been keen to research and develop autonomous vehicles. However, due to the inability of the GPS to be indoors, it is still in a state of complete autonomous travel including outdoor and indoor. In order to solve the problem of GPS, a lot of attention has been paid to a radio positioning technology for estimating the position of a user or a vehicle using a radio signal existing in an indoor space. Although wireless positioning technology is currently being commercialized and serviced, positioning accuracy is very low compared to GPS, and various types of wireless positioning technology are under development.

Wireless communication can be classified into short-range wireless communication and wide-area wireless communication. Typical examples of short-range wireless communication include Wi-Fi, Bluetooth, Zigbee, and the like. Typical examples of wide area wireless communication include 3G (3rd Generation), 4G (4th Generation), Lora And the like. LTE (Long Term Evolution) is a kind of 4G wireless communication. Bluetooth, ZigBee, etc., are not suitable for positioning because of the characteristics that occur temporarily in the indoor space according to the user's needs and disappear. Currently, Wi-Fi signals and LTE signals are known to be distributed in most rooms.

As a result, WPS (Wifi Positioning System) that performs positioning using a 2.4 GHz band Wi-Fi signal is spotlighted. A typical fingerprinting technique is a positioning method using a Wi-Fi signal. This technique divides the indoor space into a grid structure and collects signal intensity values in each unit area and builds a radio map by database. In the state where the radio map is constructed as described above, the strength of the signal received at the user location is compared with the data of the radio map to estimate the location of the user. This method has the advantage that the positioning accuracy is very high compared with the triangulation method because it collects the data reflecting the spatial characteristics of the room. As the wireless environment is good and the indoor space is finely divided and the large number of signals are collected, the positioning accuracy is improved and it can be improved up to 2 to 3 meters.

The fingerprint technique performs relatively accurate positioning when the intensity of the signal collected at the time of constructing the radio map and the intensity of the signal collected at the time of the positioning are not nearly the same. However, changes in the radio environment, such as signal interference between communication channels frequently occurring in the real world, expansion of access points, failure or occurrence of obstacles, lead to the collection of signal strengths that are different from data of the radio maps constructed in the past, This will have a serious impact on accuracy. Accordingly, various attempts have been made to improve the positioning accuracy by applying a KNN (K-Nearest Neighbor) or a particle filter to the fingerprint technique.

First of all, due to the fact that the Wi-Fi signal is distributed only in a part of the city center due to the characteristic of the short-range wireless communication, the fingerprint technique can not be used alone in the vehicle navigation system requiring self- It has inherent limitations. The LTE signals are distributed evenly throughout the indoor and outdoor areas, but there is a limit to increase the positioning accuracy because the area where the signal intensity is not changed is wide. In this way, the GPS-based map can not support the interior, and the wireless positioning map has a limitation in increasing the positioning accuracy, and also because the person collects the terrain information or the signal information of the area in which the positioning service is provided There is a problem that much time and cost are consumed in the map creation.

It is possible to accurately estimate the position of the mobile node even if the wireless environment is poor, such as signal interference between communication channels, expansion of access points, occurrence of a failure or an obstacle, or lack of the number of access points, It is possible to precisely estimate the position of the mobile node even without a separate sensor for physical landmark identification and to create a map in which very precise path information is recorded throughout the entire area, And to provide robust SLAM methods and apparatus that are robust to change. The present invention also provides a computer-readable recording medium on which a program for causing the computer to execute the SLAM method is provided. The present invention is not limited to the above-described technical problems, and another technical problem may be derived from the following description.

A SLAM (Simultaneous Localization And Mapping) method of a mobile node according to an aspect of the present invention comprises: estimating a relative position of the mobile node based on motion detection of the mobile node; Generating a variation pattern of at least one received signal strength over a plurality of time points; Correcting the estimated relative position based on a comparison between the generated signal intensity variation pattern and a signal intensity distribution in an area where the mobile node is located; And generating a map for the area by expressing the route of the area using the corrected relative position.

Wherein the at least one signal intensity change pattern is a change pattern of at least one signal strength represented by a succession of at least one signal strength received a plurality of times at a plurality of relative positions of the mobile node estimated at the plurality of time points . The SLAM method further comprises estimating a reception position of the mobile node with respect to at least one signal received at a current point in time among the plurality of points of view based on a comparison between the generated signal intensity variation pattern and the signal intensity distribution And the step of correcting the relative position may correct the estimated relative position by correcting the coordinate value of the estimated relative position by using the coordinate value of the estimated receiving position.

The SLAM method further comprises the step of extracting a portion having a pattern most similar to a pattern of variation of the generated signal intensity within the signal intensity distribution by comparing the signal intensity distribution with the variation pattern of the generated signal intensity, The step of estimating the reception position may estimate an absolute position in the region indicated by the extracted portion as a reception position of the mobile node.

Wherein the SLAM method includes a geometric shape of a surface pattern that graphs a change in at least one signal intensity according to a relative change of a position of the mobile node and a shape of a shape most similar to the surface shape within the signal intensity distribution And the step of estimating the reception position may estimate the absolute position in the region indicated by the extracted surface portion as the reception position of the mobile node.

Wherein the step of generating the at least one signal intensity change pattern maps an ID of a fixed node to a first coordinate axis of the multidimensional space, maps a relative position of the mobile node to a second coordinate axis, The pattern of the geometric surface shape can be generated by displaying dots at a point of the multidimensional space determined by mapping the intensity of a signal transmitted from any one of the fixed nodes.

Wherein the step of correcting the relative position comprises the step of estimating the relative position between the coordinate value of the estimated reception position and the coordinate value of the positioning point closest to the coordinate value of the reception position among the positioning points within the signal intensity distribution, And the positioning points may be points at which the relative position of the mobile node is measured on the movement path of the mobile node.

Wherein the step of correcting the relative position comprises: minimizing an error between a coordinate value of the estimated relative position and a coordinate value of another relative position of the mobile node based on the estimated relative position, The coordinate value of the estimated relative position can be corrected in a direction that minimizes the difference between the coordinate value and the coordinate value of the positioning point closest to the coordinate value of the reception position among the positioning points within the signal intensity distribution.

Wherein the step of correcting the relative position comprises the step of calculating a difference between a coordinate value of the arrival point that is the estimated reception position and a coordinate value of a starting point that is the closest to the coordinate value of the arrival point among the positioning points within the signal intensity distribution, And the mobile node may return from the starting point to a route of the root type to arrive at the arrival point corresponding to the same geographical position with respect to the starting point.

Wherein estimating the reception position comprises: measuring an intensity of at least one signal transmitted from the at least one fixed node; Generating a variation pattern of at least one signal strength according to a relative change of the measured at least one signal strength and a position of the mobile node from the relative position of the estimated mobile node to the plurality of time points; And estimating the reception position based on a comparison between the pattern of change of the generated at least one signal strength and the distribution of the signal strength in the area.

Wherein the step of generating a pattern of change of at least one signal strength further comprises the step of generating pattern data representing a pattern of at least one signal strength received from the at least one fixed node at the estimated relative position, Position to the pattern data for the at least one signal strength. The generating of the pattern of change of the at least one signal strength may generate the pattern data from the spatial domain data representing the respective measured signal strengths in association with the estimated relative positions.

The step of measuring the signal intensity, the step of estimating the relative position, the step of generating the pattern, and the step of correcting the relative position are repeatedly performed for each of the plurality of viewpoints, A method of mapping a coordinate value of a plurality of relative positions corrected at a plurality of points of time to a coordinate value of a corrected relative point at each point of time, The map can be created by recording the intensity of one signal.

According to another aspect of the present invention, there is provided a computer-readable recording medium recording a program for causing a computer to execute the SLAM method.

According to another aspect of the present invention, there is provided an SLAM (Simultaneous Localization And Mapping) apparatus of a mobile node, comprising: a relative positioning unit for estimating a relative position of the mobile node based on motion detection of the mobile node; A wireless positioning unit for estimating a reception position of a mobile node with respect to a signal received at a current point in time among the plurality of points of view based on a change pattern of at least one signal intensity received over a plurality of points of time; A position correcting unit for correcting the estimated relative position by correcting a coordinate value of the estimated relative position using the coordinate value of the estimated receiving position; And a map creating unit for creating a map of the area by expressing a route of the area where the mobile node is located using the corrected relative position.

Wherein the wireless positioning unit comprises: a signal processing unit for measuring an intensity of at least one signal transmitted from at least one fixed node; A pattern generating unit for generating at least one signal intensity and at least one signal intensity variation pattern according to a relative change of a position of the mobile node from the relative position of the estimated mobile node to the plurality of time points; And a reception position estimating unit estimating a reception position of the mobile node based on a comparison between the generated pattern of the at least one signal intensity and the signal intensity distribution in the area.

Wherein the SLAM apparatus further comprises a buffer for accumulating the pattern data generated by the pattern generator, wherein the pattern generator generates a pattern of at least one signal strength received from the at least one fixed node at the estimated relative position, May be accumulated in the pattern data stored in the buffer so as to generate the change pattern of the at least one signal strength.

Wherein the SLAM device further comprises a storage in which a map indicating a distribution of signal intensity in the area is stored and wherein the wireless positioning unit is configured to determine, based on a comparison of the signal intensity distribution of the signal strength and the signal intensity distribution of the map stored in the storage The reception position can be estimated.

Estimating a relative position of the mobile node based on the motion detection of the mobile node, and based on a comparison between the variation pattern of at least one signal strength received over a plurality of time points and the signal intensity distribution in an area where the mobile node is located The relative position of the mobile node is corrected and the path of the area is expressed by using the thus corrected relative position, so that the radio environment change such as signal interference between communication channels, expansion of access point, occurrence of trouble or obstacle, It is possible to accurately estimate the position of the mobile node even if the wireless environment is poor, such as a shortage of the number, and to create a map in which highly accurate route information is recorded throughout the area. Since the position estimation and the map creation of the mobile node are possible at the same time, the time and cost consumed in the map creation can be greatly reduced compared to a method in which the person directly acquires the terrain information or the signal information.

Based on a comparison between a variation pattern of at least one signal strength received over a plurality of time points and a signal intensity distribution in an area where the mobile node is located, And corrects the coordinate value of the relative position of the mobile node by using the coordinate value of the reception position estimated as described above to compensate the defects of the relative positioning and the defects of the radio positioning, It is possible to precisely estimate the position of the mobile node and create a map in which highly accurate path information is recorded throughout the region.

Unlike conventional SLAMs, it does not require a physical landmark and instead uses a similarity pattern between the pattern of change of at least one signal strength received over a plurality of time points and the corresponding pattern in the signal intensity distribution in the region where the mobile node is located The position of the mobile node can be estimated based on the degree of pattern similarity, which is much lower in complexity than the image processing of the conventional SLAM, and a map can be generated simultaneously. As a result, the SLAM algorithm can be smoothly executed without a separate sensor for physical landmark identification even in a small and light mobile node such as a smart phone.

Even in the case of estimating the position of the mobile node using a wireless signal with little change in signal intensity over a wide area, such as an LTE signal, the position of the mobile node can be estimated using at least one signal intensity change pattern received over a plurality of time points. Can be estimated accurately. Even if there is little change in the signal strength between adjacent positioning points on the movement path of the mobile node, within the movement distance corresponding to the variation pattern of the signal intensity used for the radio positioning of the present invention, This is because the strength of the LTE signal changes sufficiently to the extent that estimation can be made.

It is possible to estimate the position of the mobile node by using the LTE signal having almost no change in the signal strength between the measurement points on the movement path, and thus it is possible to create a map that can cover the outdoor space beyond the indoor space. As a result, it is possible to provide a map capable of both high-accuracy indoor positioning and outdoor positioning in the city center without the influence of high-rise buildings using LTE signals distributed widely in buildings and in urban areas. It is possible to replace GPS-based map which is widely used but can not be used for indoor positioning and the accuracy of positioning is seriously degraded in the city center, and it is possible to advance the realization of fully autonomous travel including outdoor and indoor.

1 is a configuration diagram of a wireless communication system according to an embodiment of the present invention.
Fig. 2 is a configuration diagram of the SLAM device of the mobile node 1 shown in Fig.
3 is a flow diagram of a SLAM method in accordance with an embodiment of the present invention.
4 is a detailed flowchart of step 130 shown in FIG.
5 is a view for explaining the principle of pattern formation in step 430 of FIG.
FIG. 6 is a diagram showing a three-dimensional spatial coordinate system for generating a variation pattern of signal intensity used in the SLAM algorithm of the present embodiment.
7 is a table showing the accumulation of pattern data used in the SLAM of the present embodiment.
Fig. 8 is a diagram showing an example of a map generation area according to the SLAM of the present embodiment.
9 is a diagram showing an example in which a variation pattern of signal intensity used in the SLAM of the present embodiment is generated.
10-11 are views showing examples in which the reception position of the mobile node 1 is estimated according to the SLAM algorithm of the present embodiment.
FIG. 12 is a moving locus diagram of the mobile node 1 estimated by the PDR algorithm in the target area shown in FIG.
13 is a diagram showing an example of relative positional correction by the SLAM according to the present embodiment.
14-15 are diagrams comparing SLAM performance tests for the conventional wireless positioning algorithm and the wireless positioning algorithm of the present embodiment.
16 is a diagram showing examples of maps generated by the SLAM algorithm of the present embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Hereinafter, a mobile node including a mobile phone carried by a user, a navigation system moving on a vehicle, and a mobile robot moving as a stand-alone entity, will be collectively referred to as a mobile node. It is also known as a " fixed node " including a communication device that is fixedly installed in an area and relays wireless communication of a mobile node, such as an access point (AP) of an WiFi network and a base station of an LTE network . In addition, an RF (Radio Frequency) signal transmitted from a fixed node will be briefly referred to as a " signal ".

An embodiment of the present invention to be described below relates to a SLAM (Simultaneous Localization And Mapping) method and apparatus for estimating a location of a mobile node in an unknown environment and generating a map of an unknown environment, It is possible to accurately estimate the position of the mobile node even if the radio environment such as interference, expansion of the access point, change of radio environment such as failure or obstacle occurs, or insufficient number of access points, And it is possible to precisely estimate the position of the mobile node without a separate sensor for physical landmark identification and to create a map in which highly accurate route information is recorded all over the area. ≪ / RTI > Hereinafter, the SLAM method and the SLAM apparatus will be briefly referred to as " SLAM method " and " SLAM apparatus ".

1 is a configuration diagram of a wireless communication system according to an embodiment of the present invention. Referring to FIG. 1, a wireless communication system according to the present embodiment includes a plurality of mobile nodes 1, a plurality of fixed nodes 2, and a positioning server 3. Each of the plurality of mobile nodes 1 performs wireless communication with another node through at least one type of wireless communication network while being carried by a user, mounted on a vehicle, or moving independently as an independent entity. Generally, each mobile node 1 performs wireless communication through at least two kinds of wireless communication networks, for example, WiFi network and LTE network. Each of the plurality of fixed nodes (2) relays the wireless communication of each mobile node (1) so that each mobile node (1) can access the wireless communication network and perform wireless communication with another node. The fixed node may be an access point when the mobile node 1 performs wireless communication through a Wi-Fi network, and the fixed node may be a base station when performing wireless communication through the LTE network.

The positioning server 3 stores a map provided from at least one of the plurality of mobile nodes 1. Any one of the plurality of mobile nodes 1 can create a map of the entire area to be provided with the radio positioning service and provide it to the positioning server 3. An area where the radio positioning service is to be provided can be divided and assigned to each of a plurality of mobile nodes 1. [ Each mobile node 1 can create a map of the area allocated to each of the mobile nodes 1 and provide it to the positioning server 3. The positioning server 3 can complete the map of the entire area to be provided with the radio positioning service by combining the maps provided from the plurality of mobile nodes 1. [ The positioning server 3 provides the stored map to the mobile node to be wirelessly positioned. As described below, since the map stored in the positioning server 3 includes a number of reference points and a kind of radio map in which the signal strength at each reference point is recorded, in addition to the wireless correlation based on the surface correlation map according to this embodiment, .

Fig. 2 is a configuration diagram of the SLAM device of the mobile node 1 shown in Fig. Referring to FIG. 2, the SLAM apparatus of the mobile node 1 shown in FIG. 1 includes a wireless communication unit 10, a sensor unit 20, a storage 30, a buffer 40, a relative positioning unit 50, A wireless positioning unit 60, a position correcting unit 70, and a mapping unit 80. Those skilled in the art will appreciate that such components may be implemented in hardware that provides a particular function or may be implemented in a combination of memory, processor, bus, etc. in which software that provides a particular function is written It is understandable that it is possible. The above-described components are not necessarily implemented in separate hardware, and a plurality of components may be implemented by a common hardware, for example, a combination of a processor, a memory, a bus, and the like.

As described above, the mobile node 1 may be a smart phone carried by a user, a navigation system mounted on the vehicle, or a mobile robot moving independently as an independent entity. The embodiment shown in FIG. 2 relates to a SLAM apparatus, and in addition to the configuration of the SLAM apparatus shown in FIG. 2, other configurations of a smart phone, other configurations of a navigation system, and other configurations of a mobile robot are shown in FIG. 2, Is omitted because it may be blurred. Those skilled in the art will appreciate that in the case where the mobile node 1 is implemented as a smart phone, a navigation system, a mobile robot, etc., other components besides the components shown in FIG. 2 may be added .

The wireless communication unit 10 transmits and receives signals through at least one wireless communication network. The sensor unit 20 is composed of at least one sensor for sensing the movement of the mobile node 1. The sensor unit 20 may include an acceleration sensor for measuring the acceleration of the mobile node 1 and a gyro sensor for measuring the angular velocity of the mobile node 1. [ The sensor type of the sensor unit 20 can be changed depending on what kind of device the mobile node 1 is implemented in. When the mobile node 1 is implemented as a smart phone, the sensor unit 20 may be constituted by the acceleration sensor and the gyro sensor as described above. In the case where the mobile node 1 is implemented as a navigation system mounted on a vehicle, the sensor unit 20 may be composed of an acceleration sensor and a gyro sensor as described above, and an encoder, a geomagnetic sensor, May be used.

In the storage 30, a map indicating a signal intensity distribution in an area where the mobile node 1 is located is stored. From here. The area in which the mobile node 1 is located refers to an area that is the subject of the map creation (hereinafter referred to simply as the "target area"), and may be a small area such as the interior of a certain building, It could be the same large area. When the mobile node 1 intends to create a map for an area, the mobile node 1 travels through all the routes in the area, collects the strength of signals transmitted from all the fixed nodes 2 in the area, The distribution map of the collected signal intensity can be stored in the storage 30. The signal strength distribution map stored in the storage 30 is updated by the number of complete times by repeating the SLAM method shown in FIG. 3 while the mobile node 1 completes all the paths in the target area several times. Thus, the signal intensity distribution map stored in the storage 30 is provided to the positioning server 3. [

The map stored in the mobile storage 30 is updated each time the mobile node 1 completes once all the paths in the target area and the accuracy of the map stored in the storage 30 is gradually increased by updating the map do. The accuracy of the map stored in the storage 30 is increased, but it takes a long time to complete the map. Therefore, the accuracy and the time required for the map stored in the storage 30 are bridged, It is desirable that the number of times is determined. The signal intensity distribution map may be expressed in the form of a table in which the numerical values of the signal intensities are aggregated and stored in the storage 30 and expressed in the form of a graph pattern in which dots or bars representing each signal intensity are connected, Lt; / RTI >

In addition to the wireless correlation based on the surface correlation map in the present embodiment, it is preferable to store the map in the form of an electronic table in the storage 30 in order to provide a map in the form of a general radio map that can be used for other general wireless positioning. Hereinafter, a point at which the relative position of the mobile node 1 is measured on the movement path of the mobile node 1 traveling around the target area for the purpose of making such a map is referred to simply as a " positioning point " do. In the field of radio maps of fingerprints, such positioning points may be referred to as reference points and may be referred to as nodes in the SLAM field of a mobile robot. The buffer 40 is used for accumulating the pattern data generated by the pattern generating unit 15.

The relative positioning unit 50 estimates the relative position of the mobile node 1 based on the motion detection of the mobile node 1 by the sensor unit 20. [ The relative positioning unit 50 can estimate the relative position of the mobile node 1 using a Pedestrian Dead Reckoning (PDR) algorithm or a DR (Dead Reckoning) algorithm widely known in the art to which the present embodiment belongs. The wireless positioning unit 60 is configured to transmit a signal to the mobile node 1 (hereinafter, referred to as " mobile station 1 ") for a signal received at the current one of the plurality of time points based on a change pattern of at least one signal strength received over a plurality of time points, ) Of the received signal. Here, the relative position of the mobile node 1 means a relative position with respect to the other position of the mobile node 1, and the reception position of the mobile node 1 means at least one fixed Means the current position of the mobile node 1 which receives at least one signal transmitted from the node 2. [ 2, the wireless positioning unit 60 includes a scan unit 61, a signal processing unit 62, a domain conversion unit 63, a pattern generation unit 64, a comparison unit 65, (66).

3 is a flow diagram of a SLAM method in accordance with an embodiment of the present invention. Referring to FIG. 3, the SLAM method according to the present embodiment is composed of the following steps executed by the SLAM device of the mobile node 1 shown in FIG. Hereinafter, the relative positioning unit 50 and the wireless positioning unit 60 shown in FIG. 2 will be described in detail with reference to FIG. In step 110, the scanning unit 61 of the wireless positioning unit 60 of the mobile node 1 periodically scans the frequency band of the wireless communication through the wireless communication unit 10, And receives at least one signal. The sampling rate of the time domain data to be described below is determined according to the length of the scan period of the scan unit 61. The shorter the scan period of the wireless communication unit 10, the higher the sampling rate of the time domain data described below. As a result, the accuracy of the reception position of the mobile node 1 estimated according to the present embodiment can be improved.

The ID of the fixed node 2 can be known from the signal sent from the fixed node 2 because the ID of the fixed node 2 is stored in the signal sent from the certain fixed node 2. [ If there is only one fixed node 2 within the communication coverage at the current position of the mobile node 1, the wireless communication unit 10 receives one signal from one fixed node 2 through a scanning process . The wireless communication unit 10 transmits a plurality of fixed nodes 2 from the plurality of fixed nodes 2 through the scanning process if the plurality of fixed nodes 2 exist within the communicable range at the current position of the mobile node 1. [ As shown in FIG.

Fig. 1 shows an example in which the mobile node 1 receives three signals from three fixed nodes 21, 22, and 23. It can be seen that the other fixed node 24 is located outside the communicable range of the mobile node 1. [ Since the present embodiment can be applied to an area where a wireless communication infrastructure is relatively well-equipped, the mobile node 1 receives signals of a plurality of fixed nodes 2 in most cases, but in some areas where the wireless communication infrastructure is weak, (2). In the case where no signal is received during the scanning process, since the mobile station 1 can not receive the radio positioning required by the SLAM according to the present embodiment, the mobile node 1 can not receive the signal from the fixed node 2 .

In step 120, the signal processing unit 62 of the wireless positioning unit 60 of the mobile node 1 measures the intensity of each signal received in step 110. [ The wireless positioning unit 60 of the mobile node 1 in step 130 determines the position of the mobile node 1 based on the change pattern of the at least one signal strength received from the at least one fixed node 2 around the mobile node 1 Estimates the reception position of the mobile node (1) with respect to the signal received at the current time from at least one fixed node (2). Here, the pattern of change of at least one signal strength received from at least one fixed node 2 over a plurality of points of time may include at least one signal strength .

In step 210, the relative positioning unit 50 of the mobile node 1 periodically receives the output signal of the sensor unit 20. [ The relative positioning unit 50 of the mobile node 1 calculates the movement distance and the movement direction of the mobile node 1 from the value of the output signal of the sensor unit 20 received in operation 210. [ The relative positioning unit 50 of the mobile node 1 in step 230 calculates the distance between the mobile node 1 and the previous position of the mobile node 1 based on the movement distance and the movement direction of the mobile node 1, The mobile node 1 estimates the current relative position of the mobile node 1 with respect to the previous position of the mobile node 1 by calculating the relative change of the current position of the mobile node 1. [ As described above, since the sensor type of the sensor unit 20 can be changed depending on what kind of device the mobile node 1 is implemented in, the mobile node 1 can estimate the relative position of the mobile node 1 Different navigation algorithms can be used depending on what kind of device is implemented.

For example, when the mobile node 1 is a smart phone, the relative positioning unit 50 can estimate the relative position of the mobile node 1 using the PDR algorithm. More specifically, the relative positioning unit 50 calculates the moving distance of the mobile node 1 by integrating the value of the output signal of the acceleration sensor of the sensor unit 20, and calculates the moving distance of the gyro sensor of the sensor unit 20 The moving direction of the mobile node 1 can be calculated by integrating the value of the output signal. The relative positioning unit 50 can estimate the relative position of the mobile node 1 using the DR algorithm when the mobile node 1 is mounted on the vehicle as a navigation system. For example, the relative positioning unit 50 can calculate the moving distance and the moving direction of the mobile node 1 by attaching the acceleration sensor and the gyro sensor of the sensor unit 20 to the wheel of the vehicle. Since the PDR and DR algorithms for estimating the relative position of the mobile node 1 estimate the relative position of the mobile node 1 through integration of the sensor output signal values, The error of the relative position of the mobile node 1 is accumulated.

4 is a detailed flowchart of step 130 shown in FIG. 4, in step 410, the signal processing unit 62 of the wireless positioning unit 60 of the mobile node 1 generates time domain data representing the respective signal strengths measured in step 120 in association with any one of the time points do. Here, any one of the viewpoints is used as information for distinguishing the signal received in step 110 from the previously received signal or a signal received thereafter. This point in time may be the point of receipt of each signal. The reception timing of each signal may be a time point when the signal processing unit 62 reads the time of the internal clock of the mobile node 1 at the moment when each signal is input from the wireless communication unit 10. [

In more detail, in step 410, the signal processor 62 receives the ID of the fixed node 2 that has transmitted each signal for each signal received in step 110, the reception time of each signal, To generate time domain data including at least one set of signal strengths {RSS _mn , ...} _TD that are grouped into a set of intensities. Here, RSS is an abbreviation of "Received Signal Strength", TD is an abbreviation of "Time Domain", subscript "m" indicates the order of ID of fixed node 2, "n" It indicates the order of reception time.

For example, when the SLAM method shown in FIG. 3 is repeatedly executed three times, the scan unit 61 scans surrounding signals three times. If the scan unit 61 receives only one signal transmitted from the fixed node 2 having the second ID at the time of the third signal scan, the time domain data includes only one signal strength set RSS ₂₃ . If the scan unit 61 receives the signal transmitted from the fixed node 2 having the second ID and the signal transmitted from the fixed node 2 having the third ID during the third signal scan, It will contain a set of signal strength RSS RSS ₂₃ and _33.

As described above, the time domain data may be data for distinguishing the intensity of each signal measured in operation 120 from the ID of the fixed node 2 transmitting each signal in the time domain and the reception time of each signal. Every time the SLAM method according to the present embodiment is executed, the reception timing points of a plurality of signal strength sets {RSS _mn , ...} _TD included in the time domain data generated in step 410 are all the same. Accordingly, in order to reduce the length of the time domain data, a plurality of fixed node IDs and a plurality of signal strengths may be listed and attached to signals collected at the same point in time. It will be understood by those skilled in the art that time domain data can be represented in various formats other than the above-described format.

In step 420, the domain converting unit 63 of the radio positioning unit 60 of the mobile node 1 transmits the time domain data generated in step 130 to the mobile node 1 To spatial domain data that is associated with the relative position of the target domain. At least one signal strength sets {RSS _mn, ...} are each set fixed RSS _mn each representing a set of _TD included in more In more detail, the domain converting unit 63 is the time domain data produced in step 130 By replacing the reception timing of each signal among the ID of the node 2, the reception timing of each signal, and the strength of each signal with the relative position of the mobile node 1 corresponding to the reception timing of each signal, Into at least one set of signal strengths {RSS _mn , ...} _SD that group the IDs of the mobile node 2, the mobile node 1, and the strength of each signal into one set.

Here, RSS is an abbreviation of "Received Signal Strength", SD is an abbreviation of "Space Domain", subscript "m" indicates the order of ID of the fixed node 2, "n" Represents the sequence number of the relative position of the mobile node 1 corresponding to the sequence number of the reception time. If the signal reception at step 110 and the signal reception at step 210 are synchronized and executed at substantially the same time, the relative position of the mobile node 1 corresponding to the reception timing of each signal is estimated May be the relative position of the node 1. In this case, the order of the reception timing of each signal is the order of the relative position of the mobile node 1 as it is. For example, the signal strength set RSS ₂₃ included in the spatial domain data indicates the strength of the signal received from the fixed node 2 having the second ID at the time of the third relative position estimation by the relative positioning unit 50.

If the reception of the signal at step 110 and the reception of the signal at step 210 are not synchronized, the relative position of the mobile node 1 corresponding to the reception timing of each signal is determined from the relative position And may be an estimated relative position nearest to the reception time. As described above, the time domain data is obtained by combining the ID of the fixed node 2, the reception timing of each signal, and the strength of each signal into one set, While the spatial domain data includes the ID of the fixed node 2 included in the time domain data, the relative position of the mobile node 1 estimated at the time when it is included in the time domain data, and each signal strength included in the time domain data Based data in which each signal strength is associated with the relative position of the mobile node 1 by being bundled into one set.

Since the reception timing points of the plurality of signal strength sets {RSS _mn , ...} _TD included in the time domain data generated in step 410 are all the same every time the SLAM method according to the present embodiment is executed, The relative positions of the plurality of signal strength sets {RSS _mn , ...} _SD included in the spatial domain data converted in step 420 are all the same every time the SLAM method is executed. Accordingly, in order to reduce the length of the spatial domain data, a plurality of fixed node IDs and a plurality of signal intensities may be arranged at one relative position with respect to the signals collected at the same relative position. Those skilled in the art can understand that spatial domain data can be expressed in various formats other than the above-described format.

The pattern generating unit 64 of the wireless positioning unit 60 of the mobile node 1 determines at least one signal strength measured in operation 120 and a relative position of the mobile node 1 estimated in operation 230, Thereby generating at least one signal intensity variation pattern according to the relative change in the position of the mobile node 1 over the time. In more detail, the pattern generating unit 64 generates at least one signal intensity currently received in step 110 from at least one signal strength measured in step 120 and a relative position of the mobile node 1 estimated in step 230 And a pattern of the currently received at least one signal is successively arranged in a pattern of at least one signal received before the reception of the signal in step 110 so that the position of the mobile node 1 And generates at least one signal intensity change pattern in accordance with the relative change.

The SLAM method according to the present embodiment is a method for repeatedly estimating the current position of the mobile node 1 in real time when the mobile node 1 moves to a certain path, and simultaneously generating a map for the path that the mobile node 1 has traveled While the SLAM device shown in Fig. 2 is being driven, the steps shown in Figs. 3 and 4 are continuously repeated. Since the mobile node 1 can not know the signal intensity distribution in the target area when it first passes through a certain path in the target area in order to collect the signal strength to be stored in the storage 30, , The execution of step 310 is omitted.

That is, when the mobile node 1 first passes through a certain route in the target area, the signal strength is collected at various positioning points on the route and the signal strength distribution map for the route is prepared, as many times as the number of relative positioning on the route 210, 220, 230, 110, 120, and 320 are repeatedly performed. As a result, when the mobile node 1 first completes a certain route in the target area, the relative position values of the mobile node 1 expressing the route are stored in the storage 30 as the uncorrected values. Only when the mobile node 1 completes a certain path in the target area for the first time or when it has completed a number of times or when the relative position of the mobile node 1 is corrected, the format of the map stored in the storage 30 is the same The map stored in the storage 30 is updated by the number of times of completion of the route.

In more detail, when the mobile node 1 first passes through a certain path in the target area, steps 210, 220 and 230 are performed to estimate the current relative position of the mobile node 1, and steps 110 and 120 are performed Thereby measuring the strength of at least one signal received from the fixed node 2 around the mobile node 1. [ The mobile node 1 stores the estimated relative position and the measured signal strength in the storage 30 as the signal strength at the relative position. The mobile node 1 generates a signal intensity distribution map for the path by expressing the path by the coordinate value and the signal intensity of the relative position in each of the plurality of positioning points on the path estimated or measured on the basis of the path do. That is, as the mobile node 1 passes through the plurality of positioning points on the route in turn, the storage 30 stores the coordinate value of the relative position at each positioning point in the form of a numerical value of a table or a three- The storage 30 stores a signal strength distribution map for the route.

For example, when the relative positioning unit 50 estimates the relative position of the mobile node 1 using the PDR algorithm, each positioning point on the path may be a step of the user. Since the relative position of the mobile node 1 is measured at a very short interval of about 1 meter, such as the stride of the user, the continuous array of the plurality of positioning points forms a specific path. The mobile node 1 repeats this process for all the paths in the target area, so that the signal intensity distribution map in the entire target area is completed and stored in the storage 30. According to the present embodiment, the mobile node 1 updates the coordinate values of the respective positioning points every time the route is rotated, thereby improving the accuracy of the coordinate values of the respective positioning points, The graph of the path represented by the point is optimized to approach the actual path.

5 is a view for explaining the principle of pattern formation in step 430 of FIG. Referring to FIG. 5A, the intensity of a signal transmitted from the fixed node 2 is approximately attenuated in inverse proportion to the square of the distance from the fixed node 2. When the user approaches and leaves the fixed node 2, the mobile node 1 carried by the user receives the signal of the intensity as shown in Fig. 5 (a). In general, the user does not always walk at a constant speed, but may stop temporarily while walking. While the user is temporarily stopped, as shown in FIG. 5 (b), even if the SLAM method shown in FIG. 3 is repeatedly executed many times, the intensity of the signal transmitted from the fixed node 2 is measured almost the same do. The x-axis in FIG. 5 (b) represents the time point at which the signal is measured, and the y-axis represents the signal intensity. The x-axis in FIG. 5 (c) represents the relative position (RL, Relative Location) of the mobile node 1, and the y-axis represents the signal strength.

Since the intensity of the signal transmitted from the fixed node 2 is measured every time the SLAM method shown in FIG. 3 is performed, the intensity of the signal transmitted from the fixed node 2 is continuously They are not displayed in the form of curves, and in actuality, the dots displayed at the height corresponding to the signal intensity are displayed in a continuously arranged form. When the reception timing of each signal is replaced with the relative position of the mobile node 1 by the domain converting unit 63, the signal strength of the signal generated by the pattern generating unit 64 as shown in FIG. 5 (c) The change pattern is represented by a continuous sequence of the intensities of signals received at a plurality of relative positions of the mobile node 1 estimated at a plurality of points in time. Therefore, the pattern of change of at least one signal strength generated by the pattern generating unit 64 may be a continuous pattern of the intensity of at least one signal received plural times at a plurality of relative positions of the mobile node 1 estimated at a plurality of points of time It can be said that the pattern is a change pattern of at least one signal strength represented by a sequence.

The storage 30 stores a map in the form of a table showing the distribution of the signal intensity collected in the target area according to the SLAM method according to the present embodiment, or a graph in the form of a graph showing the distribution pattern of the signal intensity. When a user repeatedly moves through the same route several times, the time required to complete the route is generally different. Even if the time required to complete the route is different when the user's travel route is the same, the various locations of the user on the route are the same. Therefore, it is not only impossible to reflect the reception timing of the signal transmitted from the fixed node 2 to the map, but it is also unnecessary. In other words, the map of the present embodiment is a map in which the ID of the fixed node 2 which transmitted a signal, the position of the point at which the signal is received, and the signal intensity distribution reflecting the intensity of the signal Is expressed.

According to the present embodiment, in order to estimate the reception position of the mobile node 1 with respect to the signal received at the current point of time, a pattern that can be matched to the signal intensity distribution map is generated. Since the positioning of the mobile node 1 is performed without knowing the position of the mobile node 1, the mobile node 1 generates time domain data representing the respective signal strengths by associating them with the reception timing of each signal, And converts the time domain data into spatial domain data in which each signal strength is associated with the relative position of the mobile node 1 corresponding to the reception timing of each signal. In order to determine the coordinates of the map, the target area of the real world around which the mobile node 1 travels is divided into a grid structure in which the distance between the scale and the scale is constant. The value of the position of a point on the map is expressed in two-dimensional coordinates with the resolution of this unit.

As shown in (c) of FIG. 5, a plurality of dots representing the strength of a plurality of signals received at a plurality of relative positions of the mobile node 1 may be concentrated have. In this case, when the maximum distance between a plurality of densely packed dots is within a distance corresponding to a coordinate resolution unit of the map, that is, a resolution unit of coordinates for representing the relative position of the mobile node 1, The same effect as that of representing one signal intensity as one dot is generated, resulting in a change pattern of signal intensity being generated. For example, if the coordinate resolution unit of the radio map is 1 meter, the effect of generating a signal intensity change pattern is generated when several dots within a meter are represented as one dot as one dot intensity Results.

In step 430, the pattern generator 64 calculates at least one signal strength received from at least one fixed node 2 at a relative position of the mobile node 1 estimated at step 230, from the spatial domain data converted at step 420 Create a pattern. The pattern of at least one signal strength generated by the pattern generator 64 in step 323 is transmitted to the at least one fixed node indicated by the spatial domain data at a relative position indicated by the spatial domain data of the movement path of the mobile node 1. [ Is a pattern of at least one signal strength that is generated by displaying at least one signal strength represented by the spatial domain data. In step 323, the pattern generator 64 has at least one set of signal intensity included in the spatial domain data received in step 310 {RSS _mn, ...}, each signal strength of the set _SD for each signal strength RSS _{mn mn} set RSS Thereby producing a pattern of at least one signal strength.

FIG. 6 is a diagram showing a three-dimensional spatial coordinate system for generating a variation pattern of signal intensity used in the SLAM algorithm of the present embodiment. 6, the x-axis of the three-dimensional space is a coordinate axis in which the IDs of the plurality of fixed nodes 2 are arranged at regular intervals, and the y-axis represents the movement path of the mobile node 1 to the relative position of the mobile node 1 The z axis is a coordinate axis obtained by dividing the measurement range of the intensity of the signal received from the plurality of fixed nodes 2 by the measurement resolution unit of the signal intensity. Those skilled in the art can understand that the information represented by each of the x-, y-, and z-axes of the three-dimensional space can be exchanged with each other. For example, the x-axis may indicate the relative position of the mobile node 1, and the y-axis may indicate the ID of the fixed node 2. [

The three-dimensional spatial coordinate system shown in FIG. 6 is based on the assumption that a user or a vehicle travel path is determined as in the case of a city center road. The signal intensity distribution map stored in the storage 30 moves along the predetermined path When constructed based on the collected signals, the signal intensity distribution of the map implies a movement path expressed in the list of signal reception positions. That is, when the pattern of change of the current signal strength of the mobile node 1 coincides with the pattern of a part of the signal intensity distribution in the map, It is located at some point in the path. In the case where the movement path of the mobile node 1 is not fixed or the height of the mobile node 1 is to be estimated in addition to the position of the mobile node 1 on the ground, Lt; / RTI > of the signal strength of < / RTI >

In order to facilitate understanding of this embodiment, ten access points corresponding to the fixed node 2 of the Wi-Fi network are listed in the x-axis of Fig. 6, and a user who is carrying the mobile node 1 on the y- It is listed as 10 meters long in meters. Therefore, the resolution unit of the relative position coordinates of the mobile node 1 is 1 meter. As described below, the variation pattern of the signal intensity compared with the signal intensity distribution map stored in the storage 30 in step 440 is a three-dimensional pattern generated in the three-dimensional space of the size shown in FIG. That is, the size of the three-dimensional space shown in FIG. 6 is compared with the signal intensity distribution map stored in the storage 30 at intervals of 10 meters with respect to the path where the mobile node 1 moved during the positioning according to the present embodiment A change pattern of the signal intensity is generated. At this time, the number of access points on the movement route of the mobile node 1 is ten. The three-dimensional space coordinate system shown in FIG. 6 is merely an example, and the number of access points and the length of the movement path of the mobile node 1 can be variously modified.

In step 430, the pattern generator 64 generates an ID of a fixed node indicated by one of the signal strength sets RSS _mn for each signal strength set RSS _mn included in the spatial domain data converted in step 420 in the x- mapping, and the branches of the three-dimensional space in which the signal strength set to the y-axis RSS _mn to map the relative location of the mobile node 1 shown, and the signal intensity set in the z-axis determined by mapping the intensity of a signal indicative of the RSS _mn And generates a graph representing the signal strength of the signal strength set RSS _mn in such a manner as to display dots on the signal strength set RSS _mn . This signal intensity graph is not an image output graph for displaying to the user but an intermediate graphical element for showing a process of generating a variation pattern of signal intensity in the form of a 3D graph used for the radio positioning of the SLAM. However, in order to facilitate understanding of the present embodiment, a signal intensity graph for each signal intensity set RSS _mn , a pattern of signal intensity at a relative position, and a pattern of change in signal intensity according to a relative position change can be visually recognized As shown in FIG.

As described above, the pattern of at least one signal strength generated by the pattern generator 64 is associated with the ID of at least one fixed node indicated by the spatial domain data and the relative position indicated by the spatial domain data, Means a pattern of at least one signal strength indicative of at least one signal strength indicative of the signal strength. Therefore, if the mobile node 1 receives only one signal, the pattern of the signal intensity at the relative position of the mobile node 1 estimated in step 230 may be one dot shape. If the mobile node 1 receives a plurality of signals, the pattern of the signal intensity at the relative position of the mobile node 1 estimated in step 230 is a straight line or a curved line expressed by a plurality of dots adjacent to each other. .

In step 430, the pattern generator 64 accumulates and stores the pattern data representing the pattern of at least one signal strength thus generated in the pattern data stored in the buffer 40. The pattern data stored in the buffer 40 is pattern data for a relative position estimated before the relative position estimation in step 230. [ By accumulating the pattern data, a change pattern of at least one signal intensity measured in step 120 is generated. The pattern data as much as necessary for generating the variation pattern of the signal intensity compared with the signal intensity distribution of the map stored in the storage () can be accumulated in the buffer 40, and a larger amount of pattern data can be accumulated. In the latter case, a variation pattern of the signal intensity is generated from a part of the pattern data accumulated in the buffer 40.

7 is a table showing the accumulation of pattern data used in the SLAM of the present embodiment. In FIG. 7 (a), pattern data accumulated in the buffer 40 is expressed in a table format. In operation 430, the pattern generator 64 may accumulate the spatial domain data in the buffer 40 in the form of a table shown in FIG. 7A. The value of "m" of "APm" in the table of FIG. 7A corresponds to the coordinate value of the x-axis of the three-dimensional space as the order of the ID of the fixed node 2 and the value of "n" of "RLn"Quot; RSS _mn " is transmitted from the fixed node 2 having the ID of " APm " corresponding to the coordinate value of the y-axis of the three-dimensional space as the order of the relative position of the node 1, Corresponds to the coordinate value of the z-axis of the three-dimensional space as the intensity of the signal received at " RLn ".

According to the pattern generating method of the pattern generating unit 64 as described above, "RSS _mn " is set on any one point of the two-dimensional plane determined by the "m" value of "APm" and the "n"Quot; RSS _mn " shown in Fig. 7 (a) forms a geometric surface in the three-dimensional space because the dot is displayed at a height corresponding to the value. As described above, in step 430, the pattern generator 64 maps the ID of one fixed node to the x-axis of the three-dimensional space, maps the relative position of the mobile node 1 to the y-axis, A change in at least one signal strength according to a relative change in the position of the mobile node 1 in such a manner that dots are displayed at a point in a three-dimensional space determined by mapping intensity of a signal transmitted from a fixed node and received at its relative position To generate a three-dimensional pattern of a geometric surface shape. A plurality of signal strength sets included in the spatial domain data accumulated in the buffer 40 may not accumulate in the buffer 40 in the form of the table of FIG. 7 (a), and may be stored in various forms Can be accumulated in the buffer (40).

Fig. 8 is a diagram showing an example of a map generation area according to the SLAM of the present embodiment. In the example of Fig. 8, the target area for mapping according to the SLAM of this embodiment is the interior of the building, and the corridor of the building has a rectangular perimeter shape. The user holds the smartphone as the mobile node 1 and starts from the starting point indicated by a dot in FIG. At this time, the mobile node 1 starts from the starting point (0, 0) and moves along the dotted line shown in FIG. In this embodiment, the position of the mobile node 1 is represented by a two-dimensional coordinate value of a two-dimensional coordinate system.

Since the mobile node 1 determines in which direction its own compass information is based on the compass information, the mobile node 1 determines whether the side of the building is tilted with respect to the north- On the two-dimensional coordinate system, as shown in Fig. 8, the plane of the building can be displayed inclined. The three-dimensional spatial coordinate system shown in FIG. 6 should be distinguished from a two-dimensional coordinate system of a map created according to the SLAM algorithm of the present embodiment as a space coordinate system for generating a signal intensity variation pattern of a three-dimensional surface shape. That is, the value of the relative position of the mobile node 1 mapped to the y-axis of the three-dimensional space shown in Fig. 6 is represented by the coordinate value of the two-dimensional coordinate system shown in Fig.

The conventional SLAM estimates the position of the mobile node based on several physical landmarks. Examples of such landmarks include doors and stairs of a building having a characteristic shape without changing its position. To identify such landmarks, sensors such as LiDAR, cameras, ultrasonic sensors and the like are required. Lidar is very high in resolution but is expensive and difficult to be applied to a small mobile node 1 such as a smart phone because of its limitation in miniaturization. It is difficult for a camera to be applied to a smart phone or the like having a low processing capability of image data as its output is image data. Ultrasonic sensors can be miniaturized, but resolution is very low and there is a limit to creating high accuracy maps. The SLAM of the present embodiment does not require landmarks as described above, unlike the conventional SLAM, and instead of the signal intensity change pattern generated by the pattern generator 64 and the signal intensity of the map stored in the storage 30 The similarity between the corresponding patterns in the distribution serves as a sort of landmark.

As described above, since the present embodiment can estimate the position of the mobile node 1 based on the pattern similarity having a very low complexity as compared with the image processing of the conventional SLAM and simultaneously generate a map, The SLAM method according to the present embodiment can be executed smoothly without a separate sensor for physical landmark identification at the node 1 as well. In particular, the map created by the SLAM method according to the present embodiment has an advantage that it can be used for radio positioning in general because it includes a signal intensity distribution corresponding to a kind of radio map in addition to the physical terrain as the route of the target area .

As described above, when the mobile node 1 makes a turn for the first time in the corridor of the building shown in FIG. 8, since the signal intensity distribution in the interior of the building can not be known, , The execution of step 310 is omitted. In the example shown in FIG. 8, since the mobile node 1 moves according to the user's walking, the relative positioning unit 50 estimates the relative position of the mobile node 1 using the PDR algorithm, The positioning point becomes each step of the user. That is, the mobile node 1 estimates the relative position of the mobile node 1 at each step of the user by performing steps 210, 220 and 230 for each step of the user, and performs steps 110 and 120, And measures the strength of at least one signal received at each step location. The mobile node 1 creates a signal intensity distribution map for the path by expressing the corridor path with the coordinate value and the signal strength of the relative position in each of the plurality of steps on the corridor path while making a turn around the corridor of the building.

As a result, the storage 30 stores a signal intensity distribution map for the route. Unlike the example shown in FIG. 8, a signal intensity distribution map can be created and stored in the storage 30 in the same manner for each path even if there are multiple paths in the target area. Since this signal intensity distribution map shows a pattern of change in signal intensity in the path when only one path is set, the map data stored in the storage 30 is created in the same format as the pattern data accumulated in the buffer 40 . As described above, if the format of the pattern data accumulated in the buffer 40 and the format of the map data stored in the storage 30 are the same, the pattern data accumulated in the buffer 40 and the map data stored in the storage 30 are matched .

In FIG. 7B, the map data stored in the storage 30 is represented in the form of a table having the same format as the format shown in FIG. 7A. The value of "m" of "APm" in the table of FIG. 7 (b) is the order number of the ID of the fixed node 2 installed in the target area and the value of "n" And "RSS _mn " is the strength of the signal transmitted from the fixed node 2 having the ID of "APm" and received at the reception position "ALn" of the mobile node 1. [ The reception position of the mobile node 1 is not a relative position based on the other position of the mobile node 1 but is a position corresponding to the pattern of the change in the signal strength generated by the pattern generation unit 64, (AL) from the viewpoint of being compared with the relative position of the mobile node 1 because it is estimated based on where it is located in the signal intensity distribution map stored in the mobile station 1. Accordingly, the reception position of the mobile node 1 is indicated as an absolute position in the table of FIG. 7 (b).

As described above, the format of the map data stored in the storage 30 is the same as the format of the pattern data accumulated in the buffer 40. [ Therefore, description of the map data stored in the storage 30 will be replaced with description of the pattern data of the buffer 40 described above. Since the signal intensity distribution map generated according to the SLAM algorithm of the present embodiment includes a kind of radio map constructed by converting the intensity of a large number of signals collected in the target area into a database, the value of " RSS _mn " Lt; / RTI >

9 is a diagram showing an example in which a variation pattern of signal intensity used in the SLAM of the present embodiment is generated. The scale of the three-dimensional spatial coordinate system shown in Fig. 9 assumes 10 times the scale of the three-dimensional spatial coordinate system shown in Fig. 6. When the user moves 20 meters along the building corridor shown in Fig. 8, According to the pattern generation method of the generation unit 64, the relative position of the mobile node 1 is estimated 20 times and a three-dimensional pattern of a surface shape corresponding to the movement distance is generated by the pattern at each of the 20 relative positions. The surface shown in FIG. 9 is formed by densities of dots of different heights. When the user moves 40 meters, 60 meters, and 80 meters, it can be seen that the three-dimensional surface pattern is extended by an amount corresponding to the movement distance. The curvature of the surface appears due to the difference in intensity between the signals transmitted from the fixed nodes 2 adjacent to each other, that is, the difference between adjacent RSS _mn .

In step 440, the comparison unit 65 of the wireless positioning unit 60 of the mobile node 1 compares the signal intensity distribution of the map stored in the storage 30 with the change pattern of at least one signal strength generated in step 430 The portion having the pattern most similar to the change pattern of at least one signal intensity generated in Step 430 is extracted within the signal intensity distribution of the map stored in the storage 30. As described above, the signal intensity distribution of the map stored in the storage 30 refers to the signal intensity distribution in the area where the mobile node 1 is located, that is, the area where the map is to be created. More specifically, the comparator 65 compares the signal intensity distribution of the map stored in the storage 30 with the three-dimensional pattern of the geometric surface type obtained by graphing the change of at least one signal intensity generated in operation 430 Dimensional pattern of the three-dimensional pattern obtained by comparing the at least one signal intensity generated in step 430 within the signal intensity distribution of the map stored in the storage 30 by comparing the three- The surface portion having a similar shape is extracted.

When the map data stored in the storage 30 is represented in the form of a table of the format shown in FIG. 7 (b), the geometric surface shape of the map signal intensity distribution is plotted from the map data stored in the storage 30 A three-dimensional pattern needs to be generated. In this case, the pattern generating unit 64 generates a three-dimensional pattern of a geometric surface type that graphs at least one signal intensity change according to a relative change of the position of the mobile node 1, And generates a geometric surface-type three-dimensional pattern of the signal intensity distribution of the map stored in the map. That is, the pattern generating unit 64 maps the IDs of one fixed node to the x-axis of the three-dimensional space with reference to the table shown in FIG. 7 (b), and the signal intensity information And stores the dots in the storage 30 in such a manner that the dots are displayed at points in the three-dimensional space determined by mapping the intensity of signals received from the fixed node in the z-axis and received at the absolute position thereof. It is possible to create a geometric surface-type three-dimensional pattern that graphs the signal intensity distribution of the map.

The pattern generating unit 64 may generate a three-dimensional pattern of a geometric surface type in which all or a part of the signal intensity distribution stored in the storage 30 is graphed. Even if the entire area of the signal intensity distribution stored in the storage 30 is plotted in the case where the size of the target area is small, there is almost no problem in the data processing of the mobile node 1. [ If the scale of the target area is large, graphing the entire signal intensity distribution stored in the storage 30 may cause a heavy load on the mobile node 1. [ The pattern generation unit 64 generates a pattern area including the current position of the mobile node 1 based on the ID of at least one fixed node 2 that has transmitted the signal scanned by the scan unit 61, 30), and a geometric surface-type three-dimensional pattern in which the signal intensity distribution of the pattern region thus set is plotted can be generated. The size of the pattern area should be appropriately set so as not to cause an error in the estimated value of the reception position of the mobile node 1 due to the size thereof without significantly affecting the load of the mobile node 1. [

As described above, according to the present embodiment, based on the surface correlation between at least one variation pattern of the signal intensity generated in step 430 and the distribution pattern of the signal intensity of the map stored in the storage 30, It determines where in the signal intensity distribution of the map stored in the storage 30 the change pattern of at least one signal strength is located. For example, such surface correlation may be calculated using a three-dimensional shape matching algorithm known to those skilled in the art to which this embodiment belongs.

In step 450, the reception position estimator 66 of the radio positioning unit 60 of the mobile node 1 calculates the signal strength distribution in the region where the mobile node 1 is located, The mobile node 1 estimates the reception position of the mobile node 1 with respect to at least one signal received in operation 110. Here, the signal intensity distribution in the area where the mobile node 1 is located is the signal intensity distribution of the map stored in the storage 30. More specifically, in step 450, the reception position estimating unit 66 estimates the absolute position in the region where the mobile node 1 is located, which is indicated by the portion extracted by the comparison in step 440, that is, Is estimated as the reception position of the mobile node (1) with respect to at least one signal received in step (110). Here, the absolute position in the region where the mobile node 1 is located refers to the position in the coordinate system of the signal intensity distribution map stored in the storage 30. [

As described above, the present embodiment differs from the prior art by using a variation pattern of at least one signal intensity according to a relative change of the position of the mobile node 1 over a plurality of points of time so far, Since the reception position of the mobile node 1 is estimated, if the length of the variation pattern of the signal strength is set to be very long, the real-time property of the positioning of the mobile node 1 may deteriorate. However, the shape similarity between the surface representing the intensity variation pattern of the signal up to the current position of the mobile node 1 and the surface representing the pattern of the signal intensity distribution of the map stored in the storage 30 is calculated by a three- So that it is possible to guarantee real-time performance of the positioning of the mobile node 1 even when the variation pattern of the signal intensity over a plurality of points of time is very long.

10-11 are views showing examples in which the reception position of the mobile node 1 is estimated according to the SLAM algorithm of the present embodiment. The scale of the three-dimensional spatial coordinate system shown in Figs. 10-11 is the same as the scale of the three-dimensional spatial coordinate system shown in Fig. 6, and is a pattern example based on the relative position of the mobile node 1 shown on the left side of Figs. Is the same as the example shown in Fig. The absolute position based pattern example shown on the right side of FIGS. 10-11 shows a map of the distribution pattern of the signal intensity for a travel path of up to 100 meters. The map stored in the storage 30 is much larger than the map shown on the right side of Figs. 10-11, but due to the limitation of one ground size, the right side of Figs. Only a portion related to the matching with the pattern shown on the left side is shown. When the user moves 20 meters, a three-dimensional surface pattern of the shape shown on the left side of FIG. 10 (a) is generated.

According to the above-described matching method based on the degree of correlation, the comparing unit 65 searches for a boldly marked portion in the pattern map shown on the right side of FIG. 10 (a). Likewise, when the user moves 40 meters, 60 meters, and 80 meters, three-dimensional surface patterns in the form of the left side of FIGS. 10 (b), 10 (c), and 10 (d) are sequentially generated. The comparator 65 sequentially searches for the areas indicated in bold in the pattern map shown on the right side of (b), (c), and (d) of FIG. 10-11. The reception position estimating unit 66 estimates the absolute position corresponding to the relative position estimated in step 230 among the plurality of absolute positions of the surface part extracted in step 440, As shown in FIG. The correspondence relationship between the relative position and the absolute position is determined from the shape matching relationship between the two surfaces. That is, the reception position estimating unit 66 estimates the absolute position of the part having the shape most similar to the shape of the relative position estimated in step 230 among the plurality of absolute positions of the surface part extracted in step 440, Position.

Various wireless positioning algorithms including KNN (K-Nearest Neighbor) algorithm, particle filter algorithm, particle filter and PDR fusion algorithm widely known as conventional wireless positioning technology commonly use only the currently received signal intensity And estimates the position of the node 1. When signal strengths other than the signal strengths collected at the time of mapping are measured due to radio environment changes such as signal interference between communication channels, access point expansion, failure or obstacle generation, The conventional radio positioning algorithm has a very high probability that the current position of the mobile node 1 is estimated to be the adjacent position other than its actual position. The larger the difference between the intensity of the signal received at the time of mapping and the intensity of the currently received signal, the greater the positioning error.

As described above, since the SLAM algorithm of the present embodiment estimates the reception position of the mobile node 1 using at least one variation pattern of the signal intensity according to the relative change of the position of the mobile node over a plurality of time points, An error in the estimated value of the reception position of the mobile node 1 hardly occurs even if a radio environment change such as a signal interference between the mobile node 1, an access point, a failure or an obstacle occurs. That is, the SLAM algorithm of the present embodiment calculates a signal strength based on the intensity of the currently received signal and the signal intensity of the past received in the path through which the mobile node 1 has been so far received, The wireless environment change at the current position of the mobile node 1 has little influence on the estimation of the current position of the mobile node 1 since the current position of the mobile node 1 is estimated.

The neighboring point of the actual position of the mobile node 1 estimated when considering the intensity of the currently received signal due to the change of the radio environment according to the conventional radio positioning algorithm is a point that is out of the path indicated by the variation pattern of the signal intensity so far do. According to the present embodiment, the change in the radio environment at the location where the mobile node 1 is currently located can not change the entire change pattern of the received signal strength in the path through which the mobile node 1 has been so far, The position of the mobile node 1 is estimated using the change pattern of at least one signal intensity according to the relative change of the position of the mobile node over a plurality of points of time so far, There is a high possibility that the actual position of the mobile node 1 receiving the current signal is estimated to be the reception position of the mobile node 1 rather than the adjacent position of the actual position of the mobile node 1 estimated according to the algorithm. Of course, if the radio environment changes continuously at various points on the movement path of the mobile node 1, a positioning error may occur, but this case hardly occurs.

In particular, the strength of a signal received from a certain fixed node 2 forms a peak when passing around the peak, and such a peak tends not to be significantly affected by radio environment change. Therefore, the length of the variation pattern of the signal intensity used in the SLAM according to the present embodiment is set to a value that the mobile node 1 has already passed through even if the currently received signal is not the adjacent portion of the peak or peak, If it is long enough to include the peak portion of the various signals on the path, it becomes very robust to radio environment changes.

As described above, the variation pattern of the signal intensity used in the SLAM of the present embodiment is a three-dimensional pattern of a geometric surface type in which graphs of changes in at least one signal intensity according to the relative change of the position of the mobile node 1 In view of the comparison between the three-dimensional pattern of the surface shape of the mobile node 1 and the three-dimensional pattern of the surface shape of the map data, the wireless environment change at the current position of the mobile node 1 corresponds to the intensity of the currently received signal And it does not affect most of the surfaces corresponding to the points other than the points of the radio environment change. That is, the change of the radio environment at the current position of the mobile node 1 has little effect on the overall shape of the surface, although it may cause some deformations of the surface shape.

Since the conventional radio positioning algorithm compares the numerical value of the currently received signal intensity with the numerical value of the signal intensity distributed in the radio map, the mobile node 1 having the numerical value closest to the numerical value of the currently received signal intensity The result is that the neighboring point of the actual position is erroneously estimated as the position of the mobile node 1. According to the SLAM algorithm of the present embodiment, since the radio environment change at the current position of the mobile node 1 hardly affects the overall shape of the surface, the shape most similar to the surface shape of the three- It is very unlikely to detect a portion of a surface different from the portion of the surface to be originally sought due to an error in the strength of the currently received signal. As described above, the positioning error of the conventional algorithm according to the comparison between the numerical value of the currently received signal intensity and the signal intensity distributed in the radio map can be originally blocked, and the positioning accuracy of the mobile node 1 is greatly improved .

LTE network base stations are installed far away from neighboring base stations so that the neighboring base stations and the relay service areas do not overlap as much as the access points of the LTE network are very expensive compared to the access points of the Wi-Fi network. As a result, the LTE signal is uniformly distributed throughout the room and outdoor, but has a characteristic that the area where the change of the signal intensity is not large is wide. As described above, since the conventional wireless positioning algorithm commonly estimates the position of the mobile node 1 using only the currently received signal strength, a change in the signal strength between the positioning points on the movement path of the mobile node 1 It is not only impossible to distinguish the positioning points by the signal intensity but also to the surrounding noise very sensitively, resulting in a large positioning error.

Even when there is almost no change in the intensity of the LTE signal between the adjacent positioning points on the movement path of the mobile node 1, the length of the variation pattern of the signal intensity used for the radio positioning of this embodiment is The strength of the LTE signal is sufficiently changed within a range of movement of the signal intensity change pattern to such an extent that accurate position estimation of the mobile node 1 is possible. Accordingly, the SLAM algorithm of the present embodiment can accurately estimate the position of the mobile node 1 even when the intensity of the LTE signal is hardly changed between the adjacent positioning points on the movement path of the mobile node 1. [

As described above, the SLAM algorithm of the present embodiment can accurately estimate the position of the mobile node 1 by using the LTE signal with almost no change in signal intensity between measurement points on the movement path, It is possible to create a map that can be displayed. As a result, the SLAM algorithm of the present embodiment can provide a map capable of both highly accurate indoor positioning and outdoor positioning in the city center without the influence of the high-rise building using the LTE signals widely distributed in the buildings and the inner city It is possible to substitute GPS based map which is most widely used in car navigation system but can not be used for indoor positioning, and the positioning accuracy is seriously degraded in the city center, and it is possible to advance the realization of fully autonomous driving including outdoor and indoor.

Although the positioning accuracy of the SLAM algorithm of the present embodiment is superior to the case of using the WiFi signal and the LTE signal in the above description, there is no limitation on the signals that can be used in the SLAM of the present embodiment, and the wireless signals such as Bluetooth, Zigbee, Laura, The positioning according to the SLAM of the present embodiment can be performed using the intensity.

In operation 310, the location corrector 70 of the mobile node 1 determines whether or not the mobile node 1 is in a state where the mobile node 1 is located, And corrects the relative position of the mobile node 1 estimated by the mobile node 1. Here, the signal intensity distribution in the area where the mobile node 1 is located is the signal intensity distribution of the map stored in the storage 30. More specifically, in step 310, the position correcting unit 70 calculates the coordinates of the relative position of the mobile node 1 estimated in step 230, using the coordinates of the reception position of the mobile node 1 estimated in step 450 The relative position estimated in step 230 is corrected. In the present embodiment, the correction of the position and the correction of the coordinate value mean that the position and the coordinate value are optimized so as to approximate the actual position or the actual coordinate value, which does not necessarily mean that the position or the coordinate value is changed.

In general, there is always an error in the position estimate because the position in the positioning field is estimated based on the probability model. For example, it is assumed that the relative positioning and the radio positioning used in the SLAM of this embodiment are also located at the mobile node 1, Location. As described above, the relative position of the mobile node 1 estimated in step 230 is very low due to error accumulation or the like. The present embodiment corrects the relative position estimated in step 230 based on the comparison between the signal intensity change pattern generated in step 430 and the distribution of the signal intensity in the area where the mobile node 1 is located, Can be reduced.

FIG. 12 is a moving locus diagram of the mobile node 1 estimated by the PDR algorithm in the target area shown in FIG. 12 (a) shows the movement trajectory of the mobile node 1 estimated by the PDR algorithm while the user holds the smartphone and starts from the starting point (0,0) and turns the building corridor two turns. From this, it can be seen that the estimated trajectory at the first turn is very different from the estimated trajectory at the second turn. Particularly, due to the PDR error accumulation, it can be seen that the estimated trajectory at the second turn is different from the actual trajectory of the mobile node 1. FIG. 12 (b) shows a state where signal strengths of signals received at the respective positioning points are connected to each other between the first moving locus and the second moving locus, and FIG. 12 (c) A variation pattern of the signal intensity generated at each positioning point between the movement locus and the second movement locus is shown in which similar positioning points are connected to each other.

In the case of (b) of FIG. 12, when the received signal is very weak or distorted at a certain position, the intensity error of the received signal at the position is very large, It can be seen that the actual positioning points between the second movement trajectories are not connected to each other and that the wrong positioning points are connected to each other. In the case of (c) of FIG. 12, even when the received signal is very weak or deformed at a certain position, the signal strength of the signal strength generated at each of the positioning points hardly affects the overall shape of the pattern Therefore, it can be seen that the actual points located at the same position between the first movement locus and the second locus locus are relatively accurately connected to each other.

Accordingly, the coordinate value of the relative position of the mobile node 1 is calibrated using the coordinate value of the reception position of the mobile node 1 estimated by the conventional wireless positioning algorithm, When the coordinate value of the relative position of the mobile node 1 is corrected using the coordinate value of the reception position of the mobile node 1 estimated by the wireless positioning algorithm of the mobile node 1, Can be corrected accurately. As a result, as the positioning accuracy with respect to the mobile node 1 is improved, it is possible to create a map in which accurate path information is recorded.

In particular, since the wireless positioning unit 60 estimates the reception position of the mobile node 1 by the wireless positioning based on the surface correlation diagram using the pattern of change of at least one signal intensity received over a plurality of time points, An error in the estimated value of the reception position of the mobile node 1 hardly occurs even if a radio environment change such as signal interference, expansion of an access point, occurrence of a failure or an obstacle occurs. In this embodiment, even if a radio environment change occurs by correcting the coordinate value of the relative position of the mobile node 1 estimated in step 230 by using the coordinate value of the reception position of the mobile node 1 estimated in step 450 The present position of the mobile node 1 can be accurately estimated through the relative position correction of the mobile node 1. [

The accuracy of such a wireless correlation based on surface correlation is very high because it utilizes a pattern of change of at least one signal intensity received over a plurality of points of time so that the signal received at the current point is very weak or deformed, A defect of the signal received at the present time point can be compensated by the change pattern of the signal received over the past several points of time. On the other hand, when the mobile node (1) is located at a point where the mobile node (1) enters a wide square from a branch point or a narrow alley that is divided into several branches, the pattern of change of signals received over various past points is, The mobile node 1 is able to move in any direction among the various roads in the case where the signal received at the current point is in an abnormal state such as when the received signal is very weak or deformed, Estimating whether to proceed is very difficult.

Accordingly, the accuracy of the wireless positioning based on the surface correlation in the present embodiment may be lowered when a path change occurs, for example, when one path is divided into several branches or a narrow alleyway enters a wide square. The relative positioning unit 40 estimates the relative position of the mobile node 1 on the basis of the movement of the mobile node 1 by the sensor unit 20 so that one length is divided into several branches, The path change does not affect the accuracy of the estimated value of the relative position of the mobile node 1.

The present embodiment uses the coordinate values of the reception position of the mobile node 1 estimated by the wireless positioning based on the surface correlation map to determine the position of the mobile node 1 estimated by the relative positioning based on the motion detection of the mobile node 1. [ By correcting the coordinate value of the relative position of the mobile node 1 by correcting the coordinates of the relative position of the mobile node 1 by correcting the relative position between the defects of the relative positioning and the defects of the radio positioning, It is possible to prevent the accuracy of the estimated value of the reception position from lowering. In other words, the present embodiment can accurately estimate the current position of the mobile node 1 always in various environmental changes such as a radio environment change and a path change through mutual complement between the defects of the relative positioning and the radio positioning, By recording the current position of the estimated mobile node 1 and the intensity of the signal received at the current position, it is possible to create a map recording highly accurate path information.

In step 310, the position correcting unit 70 calculates the coordinates of the relative position of the mobile node 1 estimated in step 230 using the coordinates of the reception position of the mobile node 1 estimated in step 450 according to the following equation (1) The relative position estimated in step 230 is corrected. Equation (1) is an equation for calculating X ^* , i.e., a new n coordinate value, which minimizes the sum of the sum of left squared terms and the sum of right squared terms of n position coordinate values corresponding to X. The solution of Equation (1) can be obtained by using a Gauss-Newton method or the like.

Assuming that the mobile node 1 is a smartphone carried by the user and that the user has performed n-1 steps from the position x ₁ of the first step, X = {x ₁ , x ₂ , x ₃ , ..., x _i , x _{i + 1} , ..., x _n } ^T. Here, superscript T means that X is a column vector in which n position coordinate values are arranged in a line. Since the moving speed and moving direction of the user are very varied, each step of the user is set to each of the positioning points on the movement path of the mobile node 1. [ The user's step can be detected by the sensor built in the smartphone. That is, the mobile node 1 estimates its relative position at each step of the user and simultaneously estimates the reception position. If the mobile node 1 is a navigation system mounted on a vehicle, a positioning point may be set at a certain distance on the movement path of the mobile node 1 instead of the user's step.

The left jegophang "x _i" is the user of the i-1-th mean the coordinates of the relative position of the i-th step to step, and, "x _{i + 1"} is (i + 1) th step to the user in the i-th step Quot; ui _{+ 1} " means a moving distance and a moving direction from the relative position of the i-th step to the relative position of the (i + 1) th step. "f (x _i , u _{i + 1} )" is a function for calculating the coordinate value of the relative position of the user's i + 1 th step by applying u _{i + 1} to the relative position coordinate value of the user's i th step it means. The coordinate system of the map that is prepared in accordance with this embodiment, if a two-dimensional coordinate _{system, "f (x i, u} i + 1)" can be expressed as the following equation (2).

Coordinate values of "x _i" in equation (2) is (x _{i, 1,} x _{i, 2),} and the coordinate values of "x _{i + 1"} is a _{(x i + 1,1, x i} + 1,2) to be. "h _{i + 1} " represents the moving direction from the relative position of the i-th step to the relative position of the (i + 1) th step, and "l _{i + 1} " represents the moving distance. " u _{i + 1} " is the moving distance and the moving direction calculated from the value of the output signal of the sensor unit 20. There is always an error in the movement distance and the movement direction, that is, u _{i + 1} , calculated from the value of the output signal of the sensor unit 20 due to the hardware error of the sensor itself, the noise introduced into the sensor, Therefore, there is always an error in the calculated value of f (x _i , u _{i + 1} ), that is, the relative position estimation value of the mobile node 1. In order to increase the accuracy of relative positioning, ). The error constraint of this relative positioning is represented by the left square term of Equation (1).

For example, " x _{i + 1} " can be calculated by applying the user's current traveling direction and the user's average pedometer to the relative position coordinate value of the i-th step in order to reduce the relative positioning error. The left square term is calculated for X = {x ₁ , x ₂ , x ₃ , ..., x _i , x _{i + 1} , ..., x _n } ^T , _{x 1, x 2, x 3} , ..., x i, the relative position of the estimated value _{x i + 1, ..., x} n} T updated when the mobile node (1) because of a sensor error to the mobile node ( 1 is prevented from being excessively deviated from the actual position. As described above, the error of the relative positioning algorithm such as PDR, DR, etc. has a characteristic in which the error of the relative position of the mobile node 1 is accumulated as the relative position estimation of the mobile node 1 is repeated. Due to these characteristics and other causes, limiting the error of f (x _i , u _{i + 1} ) to a minimum has a limitation in raising the accuracy of relative positioning. This embodiment improves the accuracy of the relative position estimate of the mobile node 1 by adding a constraint that minimizes the right squared term to the constraint of minimizing the left squared term.

"X _j " in the right square term means the current position of the mobile node 1 that has received the signal sent from the fixed node 2 and "v _j " means the current position of the mobile node 1 generated in the current position x _j of the mobile node 1 Quot; x _k " refers to a pattern of change in signal intensity, and " x _k " denotes a signal intensity generated at the current position x _j of the mobile node 1 within the signal intensity distribution in the target area among the positioning points of the map stored in the storage 30 Quot; means the absolute position of the positioning point having the pattern most similar to the change pattern of the change point. is the absolute position of the portion having the pattern most similar to the change pattern v _j of the signal intensity generated at the current position x _j of the mobile node 1 within the signal intensity distribution in the target region, and g (x _j , v _j ) .

As described above, the comparator 65 compares the degree of surface correlation between the variation pattern v _j of the signal strength generated at the current position x _j of the mobile node 1 and the signal intensity distribution in the target area, (X _j , v _j ) is a function of the wireless positioning algorithm based on the surface correlation as described above since it finds the pattern most similar to the change pattern v _j of the signal intensity generated at the position x _j of have. That is, the reception position estimating unit 66 estimates the reception position of the mobile node 1 with respect to the signal received at the current position x _j of the mobile node 1 according to g (x _j , v _j ). Thus, the calculated value of "g (x _j , v _j )" is the estimated coordinate value of the receiving position of the mobile node 1. quot; x _k " is a positioning point corresponding to the previous reception position having the coordinate value closest to the coordinate value of the reception position of the currently estimated mobile node 1, that is, the positioning point of the map stored in the storage 30 can do.

In order to satisfy the constraint that the right square term of Equation (1) must be minimized, the coordinate value of "g (x _j , v _j )" and the coordinate value of "x _k " must be the same. The coordinate values of " x _k " correspond to the coordinate values of the positioning points stored in the storage 30, that is, X = {x ₁ , x ₂ , x ₃ , ..., x _i , x _{i +} , x _n} ^T from the _{"g (x j, v j} )" of as one that has the nearest coordinate value with the calculated value or a relative position estimate of the mobile node (1) according to the relative positioning is a compensation value for the estimated value . The relative position of the mobile node 1 can not be corrected since the signal intensity distribution can not be known when the mobile node 1 first turns on a certain route in the target area, and the coordinate value of " x _k " (1).

Since the accuracy of the position estimation value of the mobile node 1 by the wireless positioning based on the surface correlation map according to the present embodiment is very high, if there is no error of the relative positioning, the position estimation value of the mobile node 1 by the relative positioning, The position estimation values of the mobile node 1 by the correlation-based radio positioning are almost the same. However, the coordinate value of "g (x _j , v _j )" and the coordinate value of "x _k " can not be the same due to the relative positioning error, typically PDR error. Therefore, the present embodiment can be applied to the case where the coordinate value of the positioning point of the map stored in the storage 30 is set to the coordinates of the positioning point of the map stored in the storage 30 in the direction of approaching the coordinate value of "g (x _j , v _j ) So that the accuracy of the coordinate value of the positioning point of the map stored in the storage 30 is improved.

In this manner, in step 310, the position correcting unit 70 calculates the coordinates of the reception position of the mobile node 1 estimated in step 450 according to Equation (1) and the measurement points in the signal intensity distribution of the map stored in the storage 30 The coordinate value of the relative position of the mobile node 1 estimated in step 230 and the signal strength of the map stored in the storage 30 in the direction of minimizing the difference of the coordinate value of the positioning point closest to the coordinate value of the reception position Correct the coordinate values of each positioning point in the distribution. The positioning points within the signal intensity distribution of the map stored in the storage 30 are points at which the relative position of the mobile node 1 on the movement path of the mobile node 1 is already measured. The mobile node 1 repeatedly updates the coordinate values of the respective positioning points by using Equation 1 while rotating a certain path in the target region several times, thereby optimizing the graph of the path represented by the plurality of positioning points to approach the actual path .

As described above, this embodiment improves the accuracy of the relative position estimate of the mobile node 1 by adding a constraint that minimizes the right-hand side square term to the constraint of minimizing the left square term. That is, in step 310, the position correcting unit 70 calculates an error between the coordinate value of the relative position estimated in step 450 and the coordinate value of another relative position of the mobile node 1 based on the relative position, The coordinates of the reception position of the mobile node 1 estimated at step 450 and the positioning point closest to the coordinate value of the reception position among the positioning points within the signal intensity distribution of the map stored in the storage 30 The coordinate value of the relative position of the mobile node 1 estimated in step 230 and the coordinate value of each positioning point in the signal intensity distribution stored in the storage 30 are corrected.

As described above, in this embodiment, the relative position estimation value of the mobile node 1 is not corrected as it is by the wireless positioning based on the surface correlation map, By adding a constraint that the error between the coordinate values of the other relative positions of the mobile node 1 based on the relative position should be minimized, it is possible to compensate the defects of the relative positioning and the defects of the radio positioning, It is possible to accurately estimate the current position of the mobile node 1 even in various environmental changes such as the position of the mobile node 1, and by accurately recording the coordinate value of the position of the mobile node 1 and the intensity of the received signal at that position, A map on which path information is recorded can be created.

13 is a diagram showing an example of relative positional correction by the SLAM according to the present embodiment. 13 (a), 13 (b), and 13 (c), when the user holds the smartphone, starts from the first positioning point, walks 11 steps, and returns to the first positioning point, the relative positioning unit 50 The step number of the mobile node 1 is displayed at the relative position of the mobile node 1 estimated for each step using the PDR algorithm. The mobile node 1 calculates the moving distance calculated from the value of the output signal of the sensor unit 20 while moving one step from the first positioning point and the relative position indicating the second positioning point with respect to the first positioning point from the moving direction . Likewise, the mobile node 1 estimates the relative position of the mobile node 1 at the 3 to 12 positioning points while moving by one step from the second positioning point.

Since the user has departed from the first positioning point and returned to the first positioning point, the first positioning point and the 12th positioning point are at the same position in the actual physical terrain. As shown in FIG. 13 (a), although the 12th positioning point should be displayed overlapping with the 1st positioning point, the 12th positioning point is not displayed at the 1st positioning point due to the accumulated PDR error, . In the conventional SLAM algorithm, a loop closure is a loop closure that starts from a certain landmark, moves along a loop-shaped path, and returns to the same landmark. When the coordinate value of the landmark estimated at the time of arrival arrives at the coordinates The coordinate value of the landmark estimated at the time of arrival is adjusted to the coordinate value of the landmark at the time of departure, thereby improving the accuracy of the map.

 As shown in FIG. 13 (b), if only the 12th positioning point is moved to the 1 positioning position, the relative positioning error at the other positioning points is ignored. 13 (c), since the PDR error accumulation, that is, the accumulation of the relative positioning errors at all the positioning points, is the main cause of the movement of the mobile node 1 in the root form, The relative positioning error of the sensor should be corrected. As described above, the conventional SLAM algorithm finds a problem of loop closure through landmark identification. In order to identify such a landmark, a sensor such as a camera, an ultrasonic sensor, or the like is required.

When the mobile node 1 starts from the starting point corresponding to the first positioning point and returns to the route of the route type and arrives at the destination corresponding to the geographically same starting point, the position correcting unit 70 calculates 1, which is the closest to the coordinate value of the arrival point of the arrival point of the mobile node 1 estimated in step 450 and the coordinate value of the arrival point of the arrival point of the map stored in the storage 30, The coordinate value of the relative position of the mobile node 1 estimated in step 230 and the coordinate value of each positioning point in the signal intensity distribution stored in the storage 30 are corrected. Here, since the coordinate value of the arrival point, which is the reception position of the mobile node 1 estimated in step 450, is the position coordinate value of the mobile node 1 estimated by the wireless positioning based on the surface correlation map, Is very high.

Accordingly, the process of adjusting the coordinate value of the destination so as to be in accordance with or near to the coordinate value of the starting point, which is the receiving position of the mobile node 1 estimated in step 450, without identifying the landmark using a sensor such as a scanner, camera, ultrasonic sensor, The error of the relative position estimation value of the destination can be greatly reduced. In this embodiment, the similarity between the signal intensity change pattern generated by the pattern generator 64 and the corresponding pattern in the signal intensity distribution of the map stored in the storage 30 plays a role of a kind of landmark instead of a physical landmark. According to the present embodiment, the accuracy of the map can be improved by solving the problem of loop closure even without a separate sensor for physical landmark identification.

14-15 are diagrams comparing SLAM performance tests for the conventional wireless positioning algorithm and the wireless positioning algorithm of the present embodiment. 8, when a user holds a smartphone and starts from a starting point (0, 0) and turns around a building corridor two times, and by using a conventional wireless positioning algorithm, The reception position of the mobile node 1 is estimated and the coordinate value of the relative position of the mobile node 1 is corrected using the reception position estimation value of the mobile node 1, A dot is displayed for each position corresponding to the coordinate value, and the dot is connected to the dot. FIG. 14 (a) shows the movement trajectory of the mobile node 1 corresponding to the array of the plurality of positioning dot dots displayed through this process.

14 (b), except that the reception position of the mobile node 1 is estimated for each positioning point using the surface correlation-based wireless positioning algorithm of the present embodiment instead of the conventional wireless positioning algorithm, The movement trajectory of the mobile node 1 is shown in the same manner as in the SLAM performance test of FIG. In particular, the tests of FIGS. 14 (a) and 14 (b) were executed in a state in which all the dozens of access points installed in the target area shown in FIG. 8 were operated. The mobile node 1 estimated by the SLAM algorithm employing the wireless positioning algorithm based on the surface correlation of the present embodiment compared with the movement locus of the mobile node 1 estimated by the SLAM algorithm employing the conventional wireless positioning algorithm It can be seen that the movement trajectory is much closer to the actual shape of the building corridor.

In FIGS. 15A, 15B and 15C, the access points installed in the target area shown in FIG. 8 are operated in the test of FIG. 14A with 50%, 20%, and 10% And the moving locus of the mobile node 1 in the case of the same progress. 15 (d), 15 (e), and 15 (f), the access points installed in the target area shown in FIG. 8 are tested in the test of FIG. 14 (b) with 50%, 20% And the moving locus of the mobile node 1 in the case of the same progress. The moving locus of the mobile node 1 estimated by the SLAM algorithm employing the conventional radio positioning algorithm is very sensitive to the number of access points and is greatly distorted as shown in Figs. 15A, 15B, and 15C Able to know.

The movement trajectory of the mobile node 1 estimated by the SLAM algorithm employing the surface correlation degree-based wireless positioning algorithm of the present embodiment from FIG. 15 (d), (e), and (f) It can be seen that it maintains a shape similar to the actual shape of a building corridor without being greatly affected by the number of access points installed in the area. As described above, according to the present embodiment, since a fewer number of access points are installed than the number of access points required for establishing a good wireless environment in a certain area, a map in which highly accurate path information is recorded even when the wireless environment is poor .

As described above, in the conventional wireless positioning algorithm, when a wireless environment change such as signal interference between communication channels, expansion of access points, occurrence of a failure or an obstacle occurs, or the number of access points is insufficient, ) Is very large. On the other hand, since the SLAM algorithm of the present embodiment estimates the reception position of the mobile node 1 by using a change pattern of at least one signal intensity according to a relative change of the position of the mobile node over a plurality of time points, An error in the estimated value of the reception position of the mobile node 1 hardly occurs even if the wireless environment changes due to a change in the radio environment such as the expansion of the access point, the occurrence of a failure or an obstacle, or a lack of the number of access points. As a result, the SLAM algorithm employing the wireless correlation algorithm based on the surface correlation of the present embodiment can obtain highly accurate path information even when the wireless environment such as a change in the number of access points or a change in radio environment occurs or the number of access points is insufficient. Map can be created.

The mapping unit 80 of the wireless positioning unit 60 of the mobile node 1 in step 320 expresses the path of the area using the relative position of the mobile node 1 corrected in step 310, Create a map. When the steps 110, 120, 130, and 310 are repeatedly performed for each of the plurality of viewpoints, in step 320, the map creator 80 displays a plurality of relative positions Coordinates of the coordinates of the mobile node 1 are recorded and stored in the storage 30, and the coordinates are mapped to coordinate values of the relative position of the mobile node 1 corrected at each point in time in the storage 30 so that at least one The ID of at least one fixed node that has transmitted the signal and the intensity of the at least one signal are recorded so as to generate a signal strength distribution map of the target area. In this way, since the position estimation and the map creation of the mobile node 1 are possible at the same time, the time and cost consumed in the map creation can be greatly reduced compared to a method in which the person directly acquires the terrain information or the signal information .

When the mobile node 1 repeatedly completes all the paths in the target area by the number of times corresponding to the target map accuracy in this embodiment, the storage 30 is provided with very high accuracy path information, signal strength information, The map on which the information is recorded is completed and stored. When the map created by the SLAM algorithm of the present embodiment includes a radio map based on a Wi-Fi signal, route information, Wi-Fi signal strength information, and access point information are recorded on the map as shown in FIG. 7, Based radio map, route information, LTE signal strength information, and base station information are recorded. If the map of the target accuracy is stored in the storage 30, the map creating unit 80 transmits the created map to the positioning server 3 via the wireless communication unit 10. [

16 is a diagram showing examples of maps generated by the SLAM algorithm of the present embodiment. FIG. 16A shows a map created by the SLAM algorithm of the present embodiment with respect to the target area shown in FIG. From Fig. 16 (a), the corridor shape in the map created by the SLAM algorithm of the present embodiment by arranging the positioning points having the relative position coordinate values of the mobile node 1 corrected by the SLAM algorithm of this embodiment is shown in Fig. 8 It can be seen that the shape of the actual corridor shown in FIG. The path graph in this map can be generated from the numerical values of the table as shown in Fig. FIG. 16 (b) shows an example of a map created by the SLAM algorithm of the present embodiment for a target area having multiple paths. The map shown in FIG. 16 (b) can be created in the same manner as the example of FIG. 16 (a) by the SLAM algorithm described above.

As shown in FIGS. 16A and 16B, in the map created by the SLAM algorithm of this embodiment, the physical terrain of the real world is simplified into a plurality of map links and a plurality of map nodes . Each map link represents a linear path through which a person or a vehicle can pass, and each map node is represented in the form of a map node indicating a point where a plurality of map links meet or a point where one map link is broken. That is, each map node represents a point that intersects a road and a length, a point that starts to divide into a plurality of branches, or a point where a road is broken. In this embodiment, each of the map links is a cluster. Each map link shows the order of each cluster. Any mobile node can perform the wireless positioning based on the surface correlation as described above by receiving the data of at least one cluster among the clusters of the map stored in the positioning server 3 from the positioning server 3, Lt; / RTI > Particularly, when the wireless positioning based on the surface correlation is executed on the map created by the SLAM algorithm according to the present embodiment, highly accurate positioning results can be obtained.

Meanwhile, the hybrid positioning method according to an embodiment of the present invention as described above can be implemented in a computer that can be created as a program executable on a processor of a computer, and the program can be recorded on a computer- have. A computer includes any type of computer capable of running programs, such as a desktop computer, a laptop computer, a smart phone, an embedded type computer, and the like. In addition, the structure of the data used in the above-described embodiment of the present invention can be recorded on a computer-readable recording medium through various means. The computer readable recording medium may be a storage medium such as a RAM, a ROM, a magnetic storage medium such as a floppy disk, a hard disk, etc., an optical reading medium such as a CD ROM, a DVD, Media.

The present invention has been described with reference to the preferred embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

1 ... mobile node
10 ... wireless communication unit 20 ... sensor unit
30 ... storage 40 ... buffer
50 ... relative positioning unit
60 ... radio positioning unit
61 ... scanning section 62 ... signal processing section
63 ... domain converting section 64 ... pattern generating section
65 ... comparing section 66 ... receiving position estimating section
70 ... position correcting unit 80 ... map creating unit
2, 21, 22, 23, 24 ... fixed node
3 ... positioning server

Claims (18)

In a SLAM (Simultaneous Localization And Mapping) method of a mobile node,
Estimating a relative position of the mobile node based on motion detection of the mobile node;
Generating a variation pattern of at least one received signal strength over a plurality of time points;
Correcting the estimated relative position based on a comparison between the generated signal intensity variation pattern and a signal intensity distribution in an area where the mobile node is located; And
And creating a map of the area by expressing the path of the area using the corrected relative position. The method according to claim 1,
Wherein the change pattern of the at least one signal strength is a change pattern of at least one signal strength represented by a consecutive order of intensity of at least one signal received a plurality of times at a plurality of relative positions of the mobile node estimated at the plurality of time points Lt; / RTI > The method according to claim 1,
Estimating a reception position of the mobile node with respect to at least one signal received at a current point in time among the plurality of points of view based on a comparison between the generated signal intensity variation pattern and the signal intensity distribution,
Wherein the step of correcting the relative position corrects the estimated relative position by correcting a coordinate value of the estimated relative position using a coordinate value of the estimated receiving position. The method of claim 3,
Further comprising the step of extracting a portion having a pattern most similar to a pattern of variation of the generated signal intensity within the signal intensity distribution by comparing the signal intensity distribution with the variation pattern of the generated signal intensity,
Wherein estimating the reception position estimates an absolute position in the region indicated by the extracted portion as a reception position of the mobile node. The method of claim 3,
A pattern of a geometric surface pattern in which a change in at least one signal intensity according to a relative change of a position of the mobile node is graphed and a surface portion having a shape most similar to the surface shape within the signal intensity distribution Further comprising the steps of:
Wherein estimating the reception position estimates an absolute position in the region indicated by the extracted surface portion as a reception position of the mobile node. 6. The method of claim 5,
Wherein the step of generating the at least one signal intensity change pattern maps an ID of a fixed node to a first coordinate axis of the multidimensional space, maps a relative position of the mobile node to a second coordinate axis, Wherein the geometric surface type pattern is generated by displaying dots at points in a multidimensional space determined by mapping intensity of a signal transmitted from any one of the fixed nodes. The method of claim 3,
Wherein the step of correcting the relative position comprises the step of estimating the relative position between the coordinate value of the estimated reception position and the coordinate value of the positioning point closest to the coordinate value of the reception position among the positioning points within the signal intensity distribution, The coordinate value of the relative position is corrected,
Wherein the positioning points are points at which the relative position of the mobile node is measured on the movement path of the mobile node. 8. The method of claim 7,
Wherein the step of correcting the relative position comprises: minimizing an error between a coordinate value of the estimated relative position and a coordinate value of another relative position of the mobile node based on the estimated relative position, And corrects the coordinate value of the estimated relative position in a direction that minimizes the difference between the coordinate value and the coordinate value of the positioning point closest to the coordinate value of the reception position among the positioning points within the signal intensity distribution. Way. 8. The method of claim 7,
Wherein the step of correcting the relative position comprises the step of calculating a difference between a coordinate value of the arrival point that is the estimated reception position and a coordinate value of a starting point that is the closest to the coordinate value of the arrival point among the positioning points within the signal intensity distribution, Correcting the coordinate value of the estimated relative position,
Wherein the mobile node departs from the starting point and travels along a route in the form of a root to arrive at the arrival point that is at the same geographical position as the starting point. The method of claim 3,
The step of estimating the reception position
Measuring an intensity of at least one signal transmitted from the at least one fixed node;
Generating a variation pattern of at least one signal strength according to a relative change of the measured at least one signal strength and a position of the mobile node from the relative position of the estimated mobile node to the plurality of time points; And
And estimating the reception position based on a comparison between the variation pattern of the generated at least one signal strength and the distribution of signal intensity in the area. 11. The method of claim 10,
Wherein the step of generating a pattern of change of at least one signal strength further comprises the step of generating pattern data representing a pattern of at least one signal strength received from the at least one fixed node at the estimated relative position, And generating a pattern of variation of the at least one signal strength by accumulating in the pattern data for the position. 12. The method of claim 11,
Wherein the step of generating the pattern of change of at least one signal strength generates the pattern data from the spatial domain data representing the respective measured signal intensities in association with the estimated relative position. The method according to claim 1,
Wherein the step of measuring the signal intensity, the step of estimating the relative position, the step of generating the pattern, and the step of correcting the relative position are repeatedly performed for each of the plurality of time points,
Wherein the step of creating the map comprises: expressing the coordinates of the plurality of relative positions corrected at the plurality of points by recording and recording the coordinate values of the relative positions, mapping the coordinates to the coordinate values of the corrected relative positions at the respective points, Wherein the map is created by recording the intensity of at least one signal received at the same point in time. A computer-readable recording medium storing a program for causing a computer to execute the method according to any one of claims 1 to 13. In a SLAM (Simultaneous Localization And Mapping) apparatus of a mobile node,
A relative positioning unit for estimating a relative position of the mobile node based on motion detection of the mobile node;
A wireless positioning unit for estimating a reception position of a mobile node with respect to a signal received at a current point in time among the plurality of points of view based on a change pattern of at least one signal intensity received over a plurality of points of time;
A position correcting unit for correcting the estimated relative position by correcting a coordinate value of the estimated relative position using the coordinate value of the estimated receiving position; And
And generating a map of the area by expressing a path of an area where the mobile node is located using the corrected relative position. 16. The method of claim 15,
The wireless positioning unit
A signal processing unit for measuring an intensity of at least one signal transmitted from at least one fixed node;
A pattern generating unit for generating at least one signal intensity and at least one signal intensity variation pattern according to a relative change of a position of the mobile node from the relative position of the estimated mobile node to the plurality of time points; And
And a reception position estimator for estimating a reception position of the mobile node based on a comparison between the generated signal intensity distribution of the at least one signal intensity and the signal intensity distribution in the region. 17. The method of claim 16,
And a buffer for accumulating the pattern data generated by the pattern generator,
Wherein the pattern generator accumulates and stores pattern data representing a pattern of at least one signal strength received from the at least one fixed node at the estimated relative position in the pattern data stored in the buffer so that the at least one signal strength And generates a change pattern of the SLAM device. 16. The method of claim 15,
Further comprising: a storage for storing a map indicative of a distribution of signal intensity in the area,
Wherein the wireless positioning unit estimates the reception position based on a comparison between the signal intensity distribution pattern of the signal intensity and the signal intensity distribution of the map stored in the storage.

Images