KR102289742B1 - Apparatus and method for compensating positioning data - Google Patents

Abstract

A calibration device is provided. The compensating device may include: a collection unit configured to collect first information and second information; an extraction unit for extracting common information of the first information and the second information; and a configuration unit configured to reconstruct the first information or reconstruct the second information by using the common information.

Description

Positioning data correction device and method

The present invention relates to an apparatus and method for correcting location data collected for various location based services (LBS).

Among LBS (Location Based Service), the positioning system that can provide high-accuracy positioning results pre-collects positioning environment information (various wireless communication signals, earth's magnetic field, topographical altitude, gravity gradient, etc.) from multiple points indoors. You can use the positioning database built by The positioning system calculates the user's location as the point at which the LBS user obtained positioning environment information using the corresponding measuring equipment in the indoor space and showed the most similar positioning environment information to that in which the user is located through comparison with the positioning database. there is.

In order to accurately calculate the user's position, the accuracy of the positioning environment information stored in the positioning database must be transferred.

There is a problem that requires a lot of time and money to collect positioning environment information in the process of building a positioning database among the entire LBS process, and various collection methods have been proposed to solve the problem.

In addition, in the process of building the positioning database, since several collectors collect positioning environment information according to the collection plan, sufficient data necessary for statistical calculation may not be obtained, or conversely, environmental information is collected by concentrating more than necessary in a specific location. there is a possibility to do If a positioning database is created based on such collected data, it may be difficult to construct a positioning database for an accurate LBS, and the quality of the LBS based on location recognition accuracy may be deteriorated.

Korean Patent Application Laid-Open No. 2018-0079901 discloses a positioning data correction apparatus and method for a location-based service and a positioning data providing system using the same.

Korea Patent Publication No. 2018-0079901

An object of the present invention is to provide a correction apparatus and a correction method for correcting positioning data collected for various location-based services (LBS) using artificial intelligence-based data clustering.

The correction method of the present invention includes a collecting step of collecting positioning environment information; a storage step of storing the collected data; performing data clustering based on the collected data; It may include; a construction step of building a positioning database using the result of the clustering.

In the correction method of the present invention, when constructing a positioning database to be used for a location-based service, the data adjacent to the positioning representative point is not directly used for generating the positioning database, but an auto-encoder-based clustering process for data adjacent to the positioning representative point is performed. It can be used later to create the positioning database.

The correction device of the present invention includes: a collection unit for collecting first information and second information; an extraction unit for extracting common information of the first information and the second information; and a configuration unit configured to reconstruct the first information or reconstruct the second information by using the common information.

According to the present invention, more accurate environmental information about a positioning point can be obtained by performing a clustering process, rather than using data adjacent to a representative point to directly generate a positioning database.

In addition, according to the present invention, erroneous collection data included in the data set is removed by using an auto encoder, a representative auto encoder of unsupervised learning among artificial intelligence, and reliable environmental information is received within a range that can be obtained. Noise can be removed and cleaned up.

The calibration apparatus and calibration method of the present invention may provide a database calibration technique for providing a high-quality LBS (Location Based Service) to a user. The correction apparatus and correction method of the present invention can be applied as a database correction technology for an alternative navigation system, particularly in a tunnel, an indoor environment, or an urban area where high-rise buildings do not reach well in a tunnel where satellite navigation signals cannot be received.

1 is a schematic diagram showing a correction device of the present invention.
2 is a schematic diagram showing a positioning database generation system of a comparative example.
3 is a schematic diagram showing an extraction unit and a constituent unit;
4 is a flowchart illustrating a correction method according to the present invention.
5 is a diagram illustrating a computing device according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those of ordinary skill in the art to which the present invention pertains can easily implement them. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

In the present specification, duplicate descriptions of the same components will be omitted.

Also, in this specification, when it is mentioned that a certain element is 'connected' or 'connected' to another element, it may be directly connected or connected to the other element, but another element in the middle. It should be understood that there may be On the other hand, in this specification, when it is mentioned that a certain element is 'directly connected' or 'directly connected' to another element, it should be understood that the other element does not exist in the middle.

In addition, the terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention.

Also, in this specification, the singular expression may include the plural expression unless the context clearly dictates otherwise.

Also in this specification, terms such as 'include' or 'have' are only intended to designate that the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist, and one or more It should be understood that the existence or addition of other features, numbers, steps, operations, components, parts or combinations thereof is not precluded in advance.

Also in this specification, the term 'and/or' includes a combination of a plurality of listed items or any of a plurality of listed items. In this specification, 'A or B' may include 'A', 'B', or 'both A and B'.

Also, in this specification, detailed descriptions of well-known functions and configurations that may obscure the gist of the present invention will be omitted.

1 is a schematic diagram showing a correction device of the present invention.

The calibration apparatus of the present invention does not use data adjacent to the positioning representative point to directly generate the positioning database 90 when constructing a database to be used for LBS (Location Based Service), but rather goes through a clustering process to provide a more accurate positioning representative point. The goal is to provide accurate and efficient LBS to users by selecting

At this time, various techniques such as principal component analysis (PCA), a K-means clustering technique, or a combination of these two techniques have been proposed among the clustering techniques established in the existing system. However, data clustering is practically very difficult in situations such as when data to be clustered are dense with each other (cluster centers are close), data is distorted or damaged, and data noise is large.

Therefore, auto-encoder-based clustering has been developed for effective clustering in the various cases mentioned above. In particular, when auto-encoder is used, some difficulties in interpretation of results, a problem with artificial neural network-based techniques, can be partially solved. When using an auto-encoder, the reason for interpretation and judgment can be explained by exploring the output characteristics while effectively reconstructing the input data while using as few components of the expression vector as possible.

The correction apparatus and correction method of the present invention may propose various means and methods for applying the auto encoder-based deep clustering to the construction of the positioning database 90 . For example, the calibration device of the present invention may include a collection unit 110 , an extraction unit 130 , and a configuration unit 150 .

The collection unit 110 may collect the first information and the second information.

The extraction unit 130 may extract common information of the first information and the second information.

The configuration unit 150 may reconstruct or correct the first information using common information. The configuration unit 150 may reconstruct or correct the second information using common information.

The first information may include positioning environment information obtained from the first means.

The second information may include positioning environment information obtained from the second means.

A clustering means 190 for clustering the first information and the second information may be provided.

The clustering means 190 may include an auto-encoder that receives an input value of a set size and outputs the same size.

The extraction unit 130 may form an encoder of an auto-encoder that compresses an input value to a small size based on a position element included in the input value.

The configuration unit 150 may form a decoder of the auto-encoder that restores the value compressed by the extraction unit 130 back to the size of the input value of the extraction unit 130 .

The auto-encoder may include an encoding model and a decoding model generated by the learning unit 170 that learns by an unsupervised learning method using a feature vector.

The collection unit 110 may collect positioning environment information from a plurality of means for a location-based service.

The collection unit 110 may include a first collection unit 111 , a second collection unit 112 , and a matching unit 119 .

The first collection unit 111 may collect a plurality of pieces of first information a from the first means. As an example, the first means may include means for receiving a wireless communication signal and measuring the strength of the received signal. The intensity information of the corresponding signal may be used to determine location information where the corresponding measuring means is located. In other words, the signal strength information may be regarded as corresponding to the position information indicating the position. The first collection unit 111 may communicate with the first means and obtain first information from the first collection unit 111 .

The second collection unit 112 may collect a plurality of pieces of second information b from the second means. As an example, the second means may include a positioning location calculation means for calculating the current location using the Earth's magnetic field, terrain altitude, gravity gradient, signal strength, and the like. The calculation result of the calculation means can be used as coordinate information as it is. The second collection unit 112 may communicate with the second means and obtain second information from the second collection unit 112 .

The matching unit 119 may match specific first information and specific second information based on a location element included in each of a plurality of pieces of first information collected from the first means and a plurality of pieces of second information collected from the second means. can The first information and the second information may include various components in addition to the location element indicating or implying the location. In this case, the matching unit 119 may match the first information and the second information using only a location element related to location determination. The specific first information and the specific second information matched by the matching unit 119 may be inputted to the extraction unit 130 together as an input set a-b.

Fig. 2 is a schematic diagram showing a system for generating a positioning database 90 in a comparative embodiment.

The positioning database 90 generation system of the comparative embodiment uses data adjacent to the representative point to directly generate the positioning database 90 . In other words, the first information collected by the first collecting unit 111 and the second information collected by the second collecting unit 112 may be directly recorded in the positioning database 90 .

If the first information and the second information are recorded in the positioning database 90 without separate correction in a tunnel or indoor environment where distortion or loss of radio signals, magnetic fields, etc. exist, the location is determined based on the positioning database 90 later. It may be difficult for LBS to be provided normally. In order to provide a normal LBS, it is preferable that the first information and the second information go through a clustering process. In this case, the extraction unit 130 and the configuration unit 150 may be additionally used to overcome the practical difficulties of the clustering process. there is.

The extraction unit 130 may extract common information using specific first information and specific second information that match each other.

3 is a schematic diagram showing an extraction unit 130 and a constituent unit 150 .

For example, the extraction unit 130 may include an input unit 131 , a sharing unit 132 , and an extraction unit 133 .

The input unit 131 may include a first input unit and a second input unit.

The first input unit may have a deep neural network for extracting a location element from the first information a. A location element included in the first information may be extracted while the first information passes through a plurality of layers provided in the deep neural network.

The second input unit may have a deep neural network for extracting the location element from the second information b. A location element included in the second information may be extracted while the second information passes through a plurality of layers provided in the deep neural network.

The sharing unit 132 may generate one shared data d1 in which the location element of the first information a and the location element of the second information b are shared.

The extraction unit 133 may extract common information d2 common to each other among the location element of the first information and the location element of the second information by using the shared data d1.

The common information d2 extracted by the extraction unit 133 may be formed with only a common part among the location element of the first information a and the location element of the second information b. A portion common to each other in the first information and the second information obtained from different sources is inevitably smaller than the size or dimension of the first information a. Similarly, the portion common to each other in the first information and the second information is inevitably smaller than the size or dimension of the second information b. Accordingly, the common information d2 may form compressed data having a smaller size or a smaller dimension than the shared data d1 in which the first information and the second information are combined.

According to the first input unit, the second input unit, the sharing unit 132 , and the extraction unit 133 , the input set ab of the specific first information and the specific second information matched with each other has a smaller size and dimension than the corresponding input set ab. A compression or encoding process of the auto-encoder compressed with common information may be completed.

The collection unit 110 may receive signal strength information corresponding to the first information from the first means for receiving the wireless communication signal and measuring the strength of the signal. The collection unit 110 may collect positioning location information corresponding to the second information from a second means for calculating a positioning location using a setting source such as a signal strength of a wireless signal, the earth's magnetic field, a terrain altitude, a gravity gradient, etc. .

The extraction unit 130 generates one shared data d1 that includes both signal strength information and positioning location information, and uses the shared data d1 to extract common common information d2 between the first information a and the second information b. there is.

In this case, the configuration unit 150 may be provided with a reconfiguration unit 151 and an output unit 152 .

The output unit 152 may include a first output unit and a second output unit.

The reconstruction unit 151 may generate one piece of reconstruction data d3 in which the common information d2 is reconstructed to the size of the shared data d1.

The first output unit may have a deep neural network that generates first output information a′ having a size following the first information by using information corresponding to the first information among the reconstruction data.

The second output unit may have a deep neural network that generates second output information b′ having a size following the second information by using information corresponding to the second information among the reconstruction data.

The common information d2 has a compressed size of the shared data d1 and may include at least a portion of a common location element between the first information and the second information. When the common information d2 is reconstructed to the size of the original shared data, the location element corresponding to the common information d2 is also surely included in the shared data, and the common location element is the first output information a' and the second output information b' may also be included.

The first output information a' including common information may be in a state in which noise and distortion included in the first information a' are removed or corrected based on the location element.

The second output information b′ including the common information may be in a state in which noise and distortion included in the second information b are removed or corrected based on the location element.

As a result, the first output information may correspond to the position-corrected first information, and the second output information may correspond to the position-corrected second information. The first output information and the second output information are stored in the positioning database 90, and may be used later in the LBS.

4 is a flowchart illustrating a correction method according to the present invention.

The correction method illustrated in FIG. 4 may be performed by the correction apparatus illustrated in FIG. 1 .

The correction method of the present invention may include a collecting step (S510), a storing step (S520), a first performing step (S530), a second performing step (S540), and a building step (S550).

In the collection step ( S510 ), positioning environment information may be collected. The positioning environment information may include various kinds of raw information that can be used for positioning, such as signal strength of a wireless communication signal obtained from various means for measuring a position, strength/direction of the Earth's magnetic field, terrain altitude, gravity gradient, and the like. The collecting step may be performed by the collecting unit 110 .

The storage step (S520) may store the collected data corresponding to the positioning environment information. The storage step may be performed by the collection unit 110 . The collection unit 110 may match and store the collected data based on the location element.

In the first performing step ( S530 ) and the second performing step ( S540 ), artificial intelligence-based data clustering may be performed based on the collected data. The first performing step may be performed by the extraction unit 130 , and the second performing step may be performed by the configuration unit 150 .

The first performing step and the second performing step may be performed in an auto-encoder method.

The extraction unit 130 performs the function of encoding to compress the input value in the auto-encoder, and the configuration unit 150 performs the function of decoding to reconstruct the compressed value encoded in the auto-encoder to the size of the previous input value. can In this case, the reference of encoding and decoding may be a position element included in the first information and the second information forming the input value. Accordingly, auto encoding is performed based on the location element, and the location element of the wrong or missing specific information can be normally corrected in an indoor environment where various kinds of distortion, interference, and noise phenomena are frequent.

The first performing step and the second performing step use a multi-modal deep auto encoder corresponding to a deep learning algorithm to perform common information existing among a plurality of types of collected data regardless of the type of the collected data. can be extracted. In the first performing step and the second performing step, the original value of the missing or distorted input value may be inferred based on common information.

The construction step ( S550 ) may build the positioning database 90 using the result of clustering. The result of clustering may include first output information a' and second output information b' corresponding to the result of the auto encoder. The configuration unit 150 may transmit the first output information and the second output information that have been corrected or reconfigured to the positioning database 90 .

The positioning database 90 included in the calibration device of the present invention or formed separately stores the first output information and the second output information provided from the configuration unit 150, and may later be provided to the LBS.

In the above correction method, when the positioning database 90 to be used for the location-based service is built, the data adjacent to the positioning representative point is not directly used to generate the positioning database 90, but an auto-encoder for data adjacent to the positioning representative point It can be used to generate the positioning database 90 after going through the clustering process based on it.

For example, the location representative point may mean a point whose location is already known.

When the LBS is implemented, the signal strength of a wireless communication signal having a characteristic similar to that of the representative positioning point may be detected when it is adjacent to the positioning representative point within a set distance range. In this case, the LBS determines the position of the current signal strength detection means as the positioning representative point.

According to the present invention, distortion, interference, and noise included in 'adjacent data' detected at a position adjacent to the positioning representative point within a set distance range can be corrected so that the 'adjacent data' normally represents the positioning representative point.

5 is a diagram illustrating a computing device according to an embodiment of the present invention. The computing device TN100 of FIG. 5 may be a device (eg, a calibration device, etc.) described herein.

In the embodiment of FIG. 5 , the computing device TN100 may include at least one processor TN110 , a transceiver device TN120 , and a memory TN130 . In addition, the computing device TN100 may further include a storage device TN140 , an input interface device TN150 , an output interface device TN160 , and the like. Components included in the computing device TN100 may be connected by a bus TN170 to communicate with each other.

The processor TN110 may execute a program command stored in at least one of the memory TN130 and the storage device TN140. The processor TN110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to an embodiment of the present invention are performed. The processor TN110 may be configured to implement procedures, functions, methods, and the like described in connection with an embodiment of the present invention. The processor TN110 may control each component of the computing device TN100 .

Each of the memory TN130 and the storage device TN140 may store various information related to the operation of the processor TN110 . Each of the memory TN130 and the storage device TN140 may be configured as at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory TN130 may include at least one of a read only memory (ROM) and a random access memory (RAM).

The transceiver TN120 may transmit or receive a wired signal or a wireless signal. The transceiver TN120 may be connected to a network to perform communication.

On the other hand, the embodiment of the present invention is not implemented only through the apparatus and/or method described so far, and a program for realizing a function corresponding to the configuration of the embodiment of the present invention or a recording medium in which the program is recorded may be implemented. And, such an implementation can be easily implemented by those skilled in the art from the description of the above-described embodiments.

Although the embodiment of the present invention has been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims are also presented. It belongs to the scope of the invention.

90...position database 110...acquisition unit
111...first collection unit 112...second collection unit
119...matching unit 130...extraction unit
131...input unit 132...share unit
133...extraction 150...constituting unit
151...Reconstruction 152...Output
170...Learning 190...Means of clustering

Claims (9)

A calibration method performed by a calibration device, comprising:
collecting the first information and the second information;
extracting common information using one piece of shared data in which the first information and the second information are shared;
generating one reconstructed data in which the common information is reconstructed to the size of the shared data;
generating first output information having a size following the first information by using information corresponding to the first information among the reconstruction data;
generating second output information having a size following the second information by using information corresponding to the second information among the reconstruction data;
A calibration method comprising
delete delete a collection unit for collecting the first information and the second information;
an extraction unit for extracting common information of the first information and the second information;
a configuration unit that reconstructs or corrects the first information or reconstructs or corrects the second information by using the common information;
The extraction unit extracts the common information by using one shared data in which the first information and the second information are shared,
A reconstruction unit, a first output unit, and a second output unit are provided,
The reconstruction unit generates one reconstruction data in which the common information is reconstructed to the size of the shared data,
The first output unit has a deep neural network that generates first output information having a size following the first information by using information corresponding to the first information among the reconstruction data,
The second output unit is a calibrating device having a deep neural network for generating second output information having a size following the second information by using information corresponding to the second information among the reconstruction data.
5. The method of claim 4,
The first information includes positioning environment information obtained from the first means,
The second information includes positioning environment information obtained from the second means,
A clustering means for clustering the first information and the second information is provided;
The clustering means includes an auto-encoder that receives an input value of a set size and outputs the same size,
The extraction unit forms an encoder of the auto-encoder that compresses the input value to a small size based on a position element included in the input value,
and the configuration unit forms a decoder of the auto-encoder that restores the value compressed by the extraction unit back to the size of the input value.
5. The method of claim 4,
The collection unit collects positioning environment information from a plurality of means for a location-based service,
The collection unit matches specific first information and specific second information based on a location element included in each of a plurality of pieces of first information collected from a first means and a plurality of pieces of second information collected from a second means, ,
The extraction unit extracts the common information by using the specific first information and the specific second information that match each other.
5. The method of claim 4,
A first input unit, a second input unit, a sharing unit, and an extraction unit are provided;
The first input unit has a deep neural network for extracting a location element from the first information,
The second input unit has a deep neural network for extracting a location element from the second information,
The sharing unit generates one shared data in which the location element of the first information and the location element of the second information are shared,
The extraction unit extracts the common information common to each other among the location element of the first information and the location element of the second information by using the shared data.
5. The method of claim 4,
The collection unit collects signal strength information corresponding to the first information from a first means for receiving a wireless communication signal and measuring the strength of the signal,
The collection unit collects positioning location information corresponding to the second information from a second means for calculating a positioning location using a setting source,
The extraction unit generates one shared data including both the signal strength information and the positioning location information, and extracts the common information common between the first information and the second information by using the shared data. .
delete

Images