Method and apparatus for positioning using radio map including virtual anchor

Abstract

A positioning method of the present invention is a method for positioning a location of a user terminal after configuring a radio map by using an actual anchor that transmits a signal and a virtual anchor that does not transmit a signal, re-estimates a location of the user terminal from a measurement vector of the user terminal generated from signals received from the actual anchor, and then re-estimates a location of the user terminal on the basis of the corresponding location, and the proximity thereof to the actual anchor and the virtual anchor.

Description

Method and device for positioning using a radio map including a virtual anchor

[0001] The present invention relates to a positioning technology, and more particularly, to a positioning technology using a virtual anchor.

[0002] Although positioning systems using GPS are widely used, the positioning error is large, so only approximate positioning is possible, and positioning is impossible in indoor spaces or tunnels where GPS signals cannot reach.

[0003] To address these issues, indoor positioning technologies leveraging wireless communication technologies such as wireless local area network (WLAN), ultra-wideband (UWB), Bluetooth Low Energy (BLE), and Zigbee are emerging. These indoor positioning technologies provide accurate indoor location information using methods such as triangulation and fingerprinting.

[0004] Indoor positioning technology estimates location using communication signal information, such as signal strength information (Received Signal Strength Indication) received from anchors such as access points or beacons installed at one or more locations.

[0005] The fingerprinting method divides the indoor space into small areas (Zones) and constructs a radio map that databases the communication signal information of signals received from each anchor measured in advance for the reference location of each area, and compares the communication signal information of signals received from each anchor measured by the user terminal that wants to determine the location with the data of the radio map to search for the results that match or are most similar to the information, thereby estimating the location of the user terminal.

[0006] In the case of the fingerprint method, there is a problem that there are certain areas (e.g., corners of areas defined by the boundaries of indoor spaces) where location determination is relatively inaccurate even when located in the same indoor space.

[0007] The purpose of the present invention is to provide a method for constructing a radio map for positioning by adding a virtual anchor in addition to an actual anchor installed in an indoor space.

[0008] Additionally, the present invention aims to provide a method and device capable of determining the location of a user with high accuracy even when a relatively small number of anchors are installed in an indoor space.

[0009] Additionally, another object of the present invention is to provide a method and device capable of determining the location of a user with relatively high location accuracy even when the user is located near a boundary in an indoor space.

[0010] According to a first aspect of the present invention, a virtual anchor-based positioning method for positioning a user terminal by a positioning device includes a radio map setting step, a position estimation step, a distance calculation step, an actual metric calculation step, a virtual metric inference step, and a position re-estimation step.

[0011] The radio map setup step is a step of setting up a radio map using the location information and measurement vectors of M actual anchors that transmit signals and the location information of N virtual anchors that do not transmit signals.

[0012] The position estimation step is a step in which the position of the user terminal is estimated using the measurement vector of signals received from actual anchors.

[0013] The distance calculation step is the step of calculating the distance between the estimated location of the user terminal and the actual anchors and virtual anchors.

[0014] The actual metric calculation step is a step of calculating a metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vector stored for the actual anchors.

[0015] The virtual metric inference step is a step of inferring a metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance between the user terminal and the virtual anchors, which is inferred from the estimated location of the user terminal and the calculated distance to the virtual anchors through the learned machine learning model.

[0016] The location re-estimation step is a step of re-estimating the location of the user terminal based on the estimated location of the user terminal and the calculated metric values ​​of actual anchors and the inferred metric values ​​of virtual anchors.

[0017] The virtual anchor-based positioning method according to the second aspect of the present invention may further include an area-related anchor selection step.

[0018] The area-related anchor selection step is a step of selecting area-related anchors surrounding an area that includes the estimated location of the user terminal.

[0019] At this time, the distance calculation step calculates distances for real anchors and virtual anchors included in the area-related anchor, the actual metric calculation step calculates metric values ​​for real anchors included in the area-related anchor, the virtual metric inference step infers metric values ​​for virtual anchors included in the area-related anchor, and the location re-estimation step can re-estimate the location of the user terminal based on the calculated metric values ​​of the estimated location of the user terminal and the real anchors included in the area-related anchor and the inferred metric values ​​of the virtual anchors included in the area-related anchor.

[0020] A virtual anchor-based positioning device according to a first aspect of the present invention includes a radio map setting unit, a position estimation unit, a distance calculation unit, an actual metric calculation unit, a virtual metric inference unit, and a position re-estimation unit.

[0021] The radio map setting unit sets up a radio map using the location information and measurement vectors of M actual anchors transmitting signals and the location information of N virtual anchors not transmitting signals.

[0022] The location estimation unit estimates the location of the user terminal using the measurement vector of signals received by the user terminal from actual anchors.

[0023] The distance calculation unit calculates the distance between the estimated location of the user terminal and the actual anchors and virtual anchors.

[0024] The actual metric calculation unit calculates the metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vector stored for the actual anchors.

[0025] The virtual metric inference unit infers a metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance between the user terminal and the virtual anchors, which is inferred from the estimated location of the user terminal and the calculated distance to the virtual anchors through a learned machine learning model.

[0026] The location re-estimation unit re-estimates the location of the user terminal based on the estimated location of the user terminal and the calculated metric values ​​of the actual anchors and the inferred metric values ​​of the virtual anchors.

[0027] The virtual anchor-based positioning device according to the second aspect of the present invention may further include an area-related anchor selection unit.

[0028] The region-related anchor selection unit selects region-related anchors surrounding the region containing the estimated location of the user terminal.

[0029] At this time, the distance calculation unit calculates the distance for the real anchors and virtual anchors included in the area-related anchor, the real metric calculation unit calculates the metric value for the real anchors included in the area-related anchor, the virtual metric inference unit infers the metric value for the virtual anchors included in the area-related anchor, and the location re-estimation unit can re-estimate the location of the user terminal based on the calculated metric value of the estimated location of the user terminal and the real anchors included in the area-related anchor and the inferred metric value of the virtual anchors included in the area-related anchor.

[0030] According to the present invention, even if a relatively small number of anchors are installed in an indoor space, the user's location can be determined with high accuracy.

[0031] Additionally, according to the present invention, the user's location can be accurately determined even if the user is located at the boundary of a defined area, i.e., a corner, in an indoor space.

[0032] Figure 1 is an example of a radio map constructed based on actual anchors.

[0033] Figure 2 is an example of a radio map with a virtual anchor added.

[0034] Figure 3 illustrates the procedure of a positioning method according to the first aspect of the present invention.

[0035] Figure 4 is a flowchart of configuring a radio map with a virtual anchor added.

[0036] Figure 5 illustrates a procedure of a positioning method according to a second aspect of the present invention.

[0037] Figure 6 is a block diagram of a positioning device according to the first aspect of the present invention.

[0038] Fig. 7 is a block diagram of a positioning device according to a second aspect of the present invention.

[0039] The aforementioned and additional aspects are concretized through embodiments described with reference to the attached drawings. It is understood that various combinations of components of each embodiment are possible within the embodiment, as long as there is no other mention or contradiction between them. Each block of the block diagram may represent a physical component in some cases, but may also represent a logical representation of a portion of the function of a single physical component or a function spanning multiple physical components. Sometimes, the entity of a block or a portion thereof may be a set of program instructions. These blocks may be implemented in whole or in part by hardware, software, or a combination thereof.

[0040] Figure 1 is an example of a radio map constructed based on actual anchors.

[0041] The radio map in Fig. 1 divides the service space into four zones based on the signal strength information (RSSI), which is one of the communication signal information of the signals received from nine anchors whose locations are known. The signal strength information (RSSI from anchor #n), which is the communication signal information of the signal received from each anchor by a user terminal located at one point in the zone, n ) can be expressed as (RSSI1, RSSI2, RSSI3, RSSI4, RSSI5, RSSI6, RSSI7, RSSI8, RSSI9). However, communication signal information is not limited to signal strength information, and may include time information, channel information, various location-related sensor information, etc.

[0042] Figure 2 is an example of a radio map with a virtual anchor added.

[0043] The radio map illustrated in Fig. 2, unlike the radio map illustrated in Fig. 1, is configured by adding 16 virtual anchors that do not transmit signals but have known locations in addition to the anchors that actually transmit signals. Fig. 2 divides the service space into a total of 16 areas by 9 actual anchors (A#1 to A#9) and 16 virtual anchors (VA#1 to VA#16). As with Fig. 1, in Fig. 2, a user terminal can construct a measurement vector based on signal strength information (RSSI), which is one of the communication signal information of the received signal.

[0044] At this time, the measurement vector can be displayed only by (RSSI1, RSSI2, RSSI3, RSSI4, RSSI5, RSSI6, RSSI7, RSSI8, RSSI9) that can be obtained from signals received from actual anchors. However, the communication signal information is not limited to signal strength information, and may include time information, channel information, various location-related sensor information, etc.

[0045] FIG. 3 illustrates a procedure of a positioning method of the present invention, and FIG. 4 is a flowchart of configuring a radio map with a virtual anchor added. According to a first aspect of the present invention, a virtual anchor-based positioning method in which a positioning device determines the position of a user terminal includes a radio map setting step, a position estimation step, a distance calculation step, an actual metric calculation step, a virtual metric inference step, and a position re-estimation step.

[0046] A positioning device is a computing device that includes the concept of software in addition to hardware. In other words, a positioning device may refer to physical hardware, but it can also refer to a program running on the hardware. A positioning device is a device that includes a processor and a memory that is connected to the processor and contains program instructions executable by the processor. The device may be a computer device that further includes a storage device, a display, a network device, an input device, etc. in addition to the processor and memory. The processor is a processor that executes program instructions that implement a program, and the memory is connected to the processor and stores program instructions executable by the processor, data to be used for calculations by the processor, and data processed by the processor.

[0047] The positioning device receives a measurement vector generated by the user terminal by receiving signals from each anchor from the user terminal, and determines the position of the user terminal.

[0048] Depending on the aspect of the invention, the positioning device may be the same device as the user terminal that seeks to locate a location within the service space. In this case, there is no need for the user terminal to transmit communication signal information regarding signals received from anchors to the positioning device.

[0049] The radio map setup step, the position estimation step, the distance calculation step, the actual metric calculation step, the virtual metric inference step, and the position re-estimation step can be implemented as a set of computer program instructions, at least part of which is executed on a processor.

[0050] The radio map setting step is a step in which a positioning device sets a radio map using the location information and measurement vectors of M actual anchors transmitting signals and the location information of N virtual anchors not transmitting signals (S1000).

[0051] A positioning device using the positioning method of the present invention constructs a radio map including not only real anchors but also virtual anchors, and the method for constructing the radio map includes the steps of receiving measurement vectors from M real anchors, storing position information and measurement vectors of the M real anchors, storing position information of N virtual anchors that do not transmit signals, and constructing a radio map including the M real anchors and N virtual anchors.

[0052] Referring to FIG. 4, a positioning device sequentially selects anchors from an anchor list in which information on anchors (including virtual anchors) installed or set in a service space is stored (S1001), determines whether the selected anchor is a real anchor or a virtual anchor (determined from stored information, S1003), and then stores the location information and the measurement vector together for the real anchor (S1007), and stores the location information for the virtual anchor (S1005) to construct a radio map. The radio map construction is repeated until there are no unselected anchors in the stored anchor list (S1009).

[0053] The number N of virtual anchors is determined based on the quality of service (QoS) of the positioning-based service provided by the positioning device. That is, the higher the QoS of the positioning-based service provided, i.e., the higher the positioning accuracy and / or precision required by the service, the more virtual anchors can be set. The number of virtual anchors appropriate for the QoS can be operated as a predetermined value depending on the service space targeted by the positioning-based service and the communication signal information used. Specifically, the positioning accuracy and / or precision required by each positioning-based service can be determined in advance and operated. The positioning device can assign a service ID to each positioning-based service provided to distinguish them, and can determine the number of virtual anchors in advance according to the positioning accuracy and / or precision required for each service ID, and store this in a table for operation. Therefore, the positioning device can determine the number of virtual anchors by referring to a lookup table that stores a predetermined number of virtual anchors according to the QoS required for each positioning-based service provided.

[0054] The positioning device stores a measurement vector composed of the location information of anchors installed in the service space in advance and the communication signal information of signals received from other anchors as a radio map, and divides the service space into multiple areas and databases them. In the example shown in Fig. 2, the measurement vector of anchor A#1 is (0, RSSI 21 , RSSI 31 , RSSI 41 , RSSI 51 , RSSI 61 , RSSI 71 , RSSI 81 , RSSI 91 ), and the measurement vector of A#2 is (RSSI 12 , 0, RSSI 32 , RSSI 42 , RSSI 52 , RSSI62 , RSSI 72 , RSSI 82 , RSSI 92 )am.

[0055] The radio map established in the present invention sets virtual anchors at known locations in addition to actual anchors that transmit signals and stores their location information. At this time, the positioning device also stores location information for the virtual anchors to construct the radio map. When constructing the radio map, the positioning device does not store measurement vectors for the virtual anchors, as virtual anchors, unlike actual anchors, do not transmit or receive signals.

[0056] The actual anchor, i.e., the transmitter, may be, but is not limited to, a base station of a mobile communication system installed at a known location, a wireless LAN base station, a WiFi AP, a UWB anchor, a BLE Beacon, etc.

[0057] The position estimation step is a step in which a positioning device estimates the position of a user terminal using a measurement vector of signals received by the user terminal from actual anchors (S1010). In the position estimation step, the positioning device can estimate the position of the user terminal within the service space from the measurement vector using a positioning technology based on linear least squares (LLS) or a positioning technology based on weighted linear least squares (LBS).

[0058] The distance calculation step is a step in which the positioning device calculates the distance between the estimated position of the user terminal and the actual anchors and virtual anchors (S1020). The positioning device calculates the distance between the position of the user terminal estimated in the position estimation step and the position information of the actual anchors constituting the radio map. In addition, the positioning device calculates the distance between the two using the position of the user terminal estimated in the position estimation step and the position information of the virtual anchors constituting the radio map. In the example of Fig. 2, the Euclidean distance between the user terminal and each of the actual anchors (A#1 to A#9) is calculated based on the estimated position of the user terminal, and the Euclidean distance between the user terminal and each of the virtual anchors (VA#1 to VA#16) is calculated based on the estimated position of the user terminal.

[0059] The actual metric calculation step is a step in which the positioning device calculates a metric value through a predefined linear or nonlinear function based on a correlation coefficient or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vectors stored for the actual anchors (S1030). The positioning device calculates a correlation coefficient or Euclidean distance using the measurement vector of the user terminal and the measurement vectors stored for the actual anchors, and processes this value through a predefined linear or nonlinear function to calculate a metric value.

[0060] The metric value is a correlation coefficient or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vector stored for the actual anchor, which is obtained by processing it through a predefined linear or nonlinear function.

[0061] When the metric value is obtained using the correlation coefficient value, the larger the correlation coefficient value, the closer the user terminal is to the anchor, and the smaller the correlation coefficient value, the farther the user terminal is to the anchor.

[0062] The virtual metric inference step is a step in which a positioning device infers a metric value through a linear or nonlinear function based on a correlation coefficient value or Euclidean distance between a user terminal and virtual anchors inferred from the estimated location of the user terminal and the calculated distance to the virtual anchors through a learned machine learning model (S1040).

[0063] The trained machine learning model uses the estimated user location and the distance from the virtual anchor as input data to infer the correlation coefficient or Euclidean distance between the user terminal and the virtual anchor. In the case of virtual anchors, the measurement vector does not exist, but the trained machine learning model infers the correlation coefficient or Euclidean distance from the distance between the user location and the virtual anchor.

[0064] The machine learning model is trained to infer the correlation coefficient or Euclidean distance between the measurement vector of the user terminal and the measurement vector of the corresponding anchor, using the estimated location of the user terminal in advance and the calculated distance from the actual anchor as input data.

[0065] That is, the machine learning model uses the estimated location of the user terminal and the calculated distance from the actual anchor as input data, and is trained using the correlation coefficient value or Euclidean distance obtained using the measurement vector of the user terminal and the measurement vector of the actual anchor as the correct answer data (ground truth) for this input data.

[0066] Learning data can be generated by collecting measurement vectors through user terminals at various locations within the service space.

[0067] There are no restrictions on the types of machine learning models.

[0068] At this time, the metric value may be a correlation coefficient value inferred through a machine learning model based on the estimated location of the user terminal and the location information of the virtual anchor, similar to the value calculated between the measurement vector of the user terminal and the measurement vector of the actual anchor, or a value calculated through a predefined linear or nonlinear function based on the Euclidean distance. When the metric value is inferred using the correlation coefficient value, the larger the correlation coefficient value, the closer the user terminal is to the anchor, and the smaller the correlation coefficient value, the farther the user terminal is to the anchor.

[0069] The position re-estimation step is a step in which the positioning device re-estimates the position of the user terminal based on the estimated position of the user terminal and the calculated metric values ​​of the actual anchors and the inferred metric values ​​of the virtual anchors (S1050).

[0070] The positioning device can re-estimate the location of the user terminal by updating the weight to be applied to each anchor location when estimating the location of the user terminal using the metric value.

[0071] That is, the positioning device can re-estimate the location of the user terminal by greatly reflecting the weight to be applied to the distance to anchors with large metric values ​​when the metric value is calculated using the correlation coefficient value.

[0072] Figure 5 illustrates a procedure of a positioning method according to a second aspect of the present invention. The virtual anchor-based positioning method according to the second aspect of the present invention may further include a region-related anchor selection step.

[0073] The positioning method according to the second aspect is a method of limiting the anchors to be used when re-estimating the location of the user terminal to anchors surrounding the area where the user terminal is located.

[0074] That is, in the example of FIG. 2, if the user terminal is estimated to be located within Zone#1, the A#1 anchor, the VA#1 anchor, the VA#2 anchor, and the VA#16 anchor can be used to re-estimate the location of the user terminal.

[0075] The region-related anchor selection step may be implemented as a set of computer program instructions, at least some of whose functionality is executed on a processor.

[0076] The area-related anchor selection step is a step in which the positioning device selects an area-related anchor surrounding an area including the estimated location of the user terminal based on a radio map (S2020).

[0077] In the second aspect, the positioning device calculates a distance between the user terminal and the selected area-related anchor (S2030), the positioning device calculates a metric value through a predefined linear or nonlinear function based on a correlation coefficient value or a Euclidean distance obtained between the measurement vector of the user terminal and the stored measurement vectors for the actual anchors belonging to the area-related anchor (S2040), and the positioning device infers a metric value through a linear or nonlinear function based on a correlation coefficient value or a Euclidean distance between the user terminal and the virtual anchors belonging to the selected area-related anchors, which is inferred from the calculated distance between the estimated location of the user terminal and the virtual anchors belonging to the selected area-related anchors through a learned machine learning model (S2050). Then, the position of the user terminal can be re-estimated based on this (S2060). Accordingly, the positioning device can calculate distances for real anchors and virtual anchors included in the area-related anchors in the distance calculation step, calculate metric values ​​for real anchors included in the area-related anchors in the actual metric calculation step, infer metric values ​​for virtual anchors included in the area-related anchors in the virtual metric inference step, and re-estimate the location of the user terminal based on the calculated metric values ​​of the estimated location of the user terminal and the real anchors included in the area-related anchors and the inferred metric values ​​of the virtual anchors included in the area-related anchors in the location re-estimation step.

[0078] The signal information constituting the measurement vector may be one of positioning information including RSSI (Received Signal Strength Indicator), RSRP (Reference Signal Received Power), RSRQ (Reference Signal Received Quality), AOA (Angle Of Arrival), TOF (Time Of Flight), TOA (Time Of Arrival), TDOA (Time Difference Of Arrival), CIR (Channel Impulse Response), CFR (Channel Frequency Response), etc.

[0079] Each positioning information is well known, so a detailed explanation is omitted.

[0080] Figure 6 is a block diagram of a positioning device according to a first aspect of the present invention. The virtual anchor-based positioning device (10) according to the first aspect of the present invention includes a radio map setting unit (110), a position estimation unit (120), a distance calculation unit (130), an actual metric calculation unit (140), a virtual metric inference unit (150), and a position re-estimation unit (160).

[0081] The positioning device (10) is a computing device that includes a software concept in addition to hardware. That is, the positioning device (10) may refer to physical hardware, but may also refer to a program running on the hardware. The positioning device (10) is a device that includes a processor and a memory that is connected to the processor and includes program instructions executable by the processor. The device may be a computer device that further includes a storage device, a display, a network device, an input device, etc. in addition to the processor and the memory. The processor is a processor that executes program instructions that implement a program, and the memory is connected to the processor and stores program instructions executable by the processor, data to be used by the processor in operations, data processed by the processor, etc.

[0082] The positioning device (10) receives a measurement vector generated by the user terminal by receiving signals from each anchor from the user terminal, and determines the position of the user terminal.

[0083] Depending on the aspect of the invention, the positioning device (10) may be the same device as the user terminal that seeks to locate a location within the service space. In this case, there is no need for the user terminal to transmit communication signal information of signals received from anchors to the positioning device (10).

[0084] The radio map setting unit (110), the location estimation unit (120), the distance calculation unit (130), the actual metric calculation unit (140), the virtual metric inference unit (150), and the location re-estimation unit (160) can be implemented as a computer program instruction set, at least part of which is executed on a processor.

[0085] The radio map setting unit (110) sets a radio map using the location information and measurement vectors of M actual anchors transmitting signals and the location information of N virtual anchors not transmitting signals.

[0086] The radio map set in the present invention sets virtual anchors at known locations in addition to actual anchors that transmit signals and stores their location information. At this time, the radio map setting unit (110) also stores location information for the virtual anchors to configure the radio map. When configuring the radio map, the radio map setting unit (110) does not store measurement vectors for the virtual anchors because, unlike actual anchors, virtual anchors do not transmit or receive signals.

[0087] The number N of virtual anchors is determined based on the quality of service (QoS) of the positioning-based service provided by the positioning device (10). That is, the higher the quality of service (QoS) of the positioning-based service provided, i.e., the higher the positioning accuracy and / or precision required by the service, the more virtual anchors can be set. The number of virtual anchors appropriate for the quality of service (QoS) can be operated as a predetermined value depending on the service space that is the target of the positioning-based service and the communication signal information used. Specifically, the positioning accuracy and / or precision required by each positioning-based service can be determined in advance and operated. The positioning device (10) can assign a service ID to each positioning-based service provided to distinguish them, and can determine the number of virtual anchors in advance according to the positioning accuracy and / or precision required for each service ID, store this in a table, and operate it. Accordingly, the positioning device (10) can determine the number of virtual anchors by referring to a lookup table in which a predetermined number of virtual anchors is stored according to the quality of service (QoS) required for each positioning-based service provided.

[0088] The location estimation unit (120) estimates the location of the user terminal using the measurement vector of the signal received by the user terminal from the actual anchors. The location estimation unit (120) can estimate the location of the user terminal within the service space from the measurement vector using a linear least squares (LLS)-based location positioning technology or a weighted linear least squares (WLLS)-based location positioning technology.

[0089] The distance calculation unit (130) calculates the distance between the estimated location of the user terminal and the actual anchors and virtual anchors. The distance calculation unit (130) calculates the distance between the estimated location of the user terminal and the location information of the actual anchors that make up the radio map. In addition, the distance calculation unit (130) calculates the distance between the estimated location of the user terminal and the location information of the virtual anchors that make up the radio map.

[0090] The actual metric calculation unit (140) calculates a metric value through a predefined linear or nonlinear function based on the correlation coefficient value or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vector stored for the actual anchor. The actual metric calculation unit (140) calculates a correlation coefficient value or Euclidean distance using the measurement vector of the user terminal and the measurement vector stored for the actual anchors, and processes this value through a predefined linear function or nonlinear function to calculate a metric value.

[0091] The metric value is a correlation coefficient or Euclidean distance obtained between the measurement vector of the user terminal and the measurement vector stored for the actual anchor, which is obtained by processing it through a predefined linear or nonlinear function.

[0092] When the metric value is obtained using the correlation coefficient value, the larger the correlation coefficient value, the closer the user terminal is to the anchor, and the smaller the correlation coefficient value, the farther the user terminal is to the anchor.

[0093] The virtual metric inference unit (150) infers a metric value through a linear or nonlinear function based on the correlation coefficient value or Euclidean distance between the user terminal and the virtual anchors, which is inferred from the estimated location of the user terminal and the calculated distance from the virtual anchors through a learned machine learning model.

[0094] The trained machine learning model uses the estimated user location and the distance from the virtual anchor as input data to infer the correlation coefficient or Euclidean distance between the user terminal and the virtual anchor. In the case of virtual anchors, the measurement vector does not exist, but the trained machine learning model infers the correlation coefficient or Euclidean distance from the distance between the user location and the virtual anchor.

[0095] The machine learning model is trained to infer the correlation coefficient or Euclidean distance between the measurement vector of the user terminal and the measurement vector of the corresponding anchor, using the estimated location of the user terminal in advance and the calculated distance from the actual anchor as input data.

[0096] That is, the machine learning model uses the estimated location of the user terminal and the calculated distance from the actual anchor as input data, and is trained using the correlation coefficient value or Euclidean distance obtained using the measurement vector of the user terminal and the measurement vector of the actual anchor as the correct answer data (ground truth) for this input data.

[0097] Learning data can be generated by collecting measurement vectors through user terminals at various locations within the service space.

[0098] There are no restrictions on the types of machine learning models.

[0099] At this time, the metric value may be a correlation coefficient value inferred through a machine learning model based on the estimated location of the user terminal and the location information of the virtual anchor, similar to the value calculated between the measurement vector of the user terminal and the measurement vector of the actual anchor, or a value calculated through a predefined linear or nonlinear function based on the Euclidean distance. When the metric value is inferred using the correlation coefficient value, the larger the correlation coefficient value, the closer the user terminal is to the anchor, and the smaller the correlation coefficient value, the farther the user terminal is to the anchor.

[0100] The location re-estimation unit (160) re-estimates the location of the user terminal based on the estimated location of the user terminal and the calculated metric values ​​of the actual anchors and the inferred metric values ​​of the virtual anchors.

[0101] The location re-estimation unit (160) can re-estimate the location of the user terminal by updating the weight to be applied to each anchor when estimating the location using the measurement vector of the user terminal using the metric value.

[0102] That is, when the location re-estimation unit (160) calculates the metric value using the correlation coefficient value, it can re-estimate the location of the user terminal by increasing the weight to be applied to the distance to anchors with large metric values.

[0103] Fig. 7 is a block diagram of a positioning device according to a second aspect of the present invention. The virtual anchor-based positioning device (10) according to the second aspect of the present invention may further include an area-related anchor selection unit (170).

[0104] The positioning device (10) according to the second aspect can limit the anchors to be used when re-estimating the position of the user terminal to anchors surrounding the area where the user terminal is located.

[0105] That is, in the example of FIG. 2, if the user terminal is estimated to be located within Zone#1, the A#1 anchor, the VA#1 anchor, the VA#2 anchor, and the VA#16 anchor can be used to re-estimate the location of the user terminal.

[0106] The domain-related anchor selection unit (170) may be implemented as a set of computer program instructions, at least some of which are executed on a processor.

[0107] The area-related anchor selection unit (170) selects an area-related anchor surrounding an area including the estimated location of the user terminal based on the radio map.

[0108] In the second aspect, the distance to the selected area-related anchor is calculated, and the metric value is calculated or inferred, and then the location of the user terminal can be re-estimated based on this. Accordingly, the distance calculation unit (130) calculates the distance to the real anchor and virtual anchors included in the area-related anchor, the real metric calculation unit (140) calculates the metric value for the real anchors included in the area-related anchor, the virtual metric inference unit (150) infers the metric value for the virtual anchors included in the area-related anchor, and the location re-estimation unit (160) can re-estimate the location of the user terminal based on the calculated metric value of the estimated location of the user terminal and the real anchors included in the area-related anchor and the inferred metric value of the virtual anchors included in the area-related anchor.

[0109] The signal information constituting the measurement vector may be one of positioning information including RSSI (Received Signal Strength Indicator), RSRP (Reference Signal Received Power), RSRQ (Reference Signal Received Quality), AOA (Angle Of Arrival), TOF (Time Of Flight), TOA (Time Of Arrival), TDOA (Time Difference Of Arrival), CIR (Channel Impulse Response), CFR (Channel Frequency Response), etc.

[0110] Each positioning information is well known, so a detailed explanation is omitted.

[0111] While the present invention has been described above with reference to the accompanying drawings and examples, it is not limited thereto and should be construed to encompass various modifications that would be readily apparent to those skilled in the art. The scope of the patent claims is intended to encompass such modifications.

Claims

1. In a method for a positioning device to determine the location of a user terminal, A radio map setting step for setting a radio map using the location information and measurement vectors of M real anchors transmitting signals and the location information of N virtual anchors not transmitting signals; A position estimation step for estimating the position of a user terminal by using a measurement vector of signals received from actual anchors; A distance calculation step for calculating the distance between the estimated location of the user terminal and the actual anchors and virtual anchors; An actual metric calculation step for calculating a metric value through a linear or nonlinear function based on a correlation coefficient or Euclidean distance obtained between a measurement vector of a user terminal and a measurement vector stored for actual anchors; A virtual metric inference step for inferring a metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance between the user terminal and virtual anchors inferred from the estimated location of the user terminal and the calculated distance to the virtual anchor through the learned machine learning model; A position re-estimation step for re-estimating the position of the user terminal based on the estimated position of the user terminal and the calculated metric values ​​of the actual anchors and the inferred metric values ​​of the virtual anchors; A method for positioning using a radio map including a virtual anchor, the method comprising:

2. In paragraph 1, A region-related anchor selection step for selecting region-related anchors surrounding an area containing an estimated location of a user terminal; Including more, A positioning method using a radio map including virtual anchors, wherein the distance calculation step calculates distances for real anchors and virtual anchors included in area-related anchors, the real metric calculation step calculates metric values ​​for real anchors included in area-related anchors, the virtual metric inference step infers metric values ​​for virtual anchors included in area-related anchors, and the location re-estimation step re-estimates the location of the user terminal based on the calculated metric values ​​of the estimated location of the user terminal and the real anchors included in the area-related anchors and the inferred metric values ​​of the virtual anchors included in the area-related anchors.

3. In paragraph 1, A positioning method using a radio map including a virtual anchor, wherein a machine learning model is trained to calculate a correlation coefficient or Euclidean distance between a measurement vector of a user terminal and a measurement vector of the corresponding anchor, using the estimated position of the user terminal and the calculated distance from the actual anchor as input data.

4. In paragraph 1, A positioning method using a radio map including virtual anchors, wherein the number of virtual anchors is determined based on the quality of service (QoS) provided by the positioning device.

5. In paragraph 1, A positioning method using a radio map including a virtual anchor, wherein the signal information constituting the measurement vector is one of positioning information including RSSI, RSRP, RSRQ, AOA, TOF, TOA, TDOA, CIR, and CFR.

6. A radio map setting unit that sets a radio map using the location information and measurement vectors of M actual anchors transmitting signals and the location information of N virtual anchors not transmitting signals; A location estimation unit that estimates the location of a user terminal by using a measurement vector of signals received from actual anchors; A distance calculation unit that calculates the distance between the estimated location of the user terminal and the actual anchors and virtual anchors; An actual metric calculation unit that calculates a metric value through a linear or nonlinear function based on a correlation coefficient or Euclidean distance obtained between a measurement vector of a user terminal and a measurement vector stored for actual anchors; A virtual metric inference unit that infers a metric value through a linear or nonlinear function based on the correlation coefficient or Euclidean distance between the user terminal and virtual anchors inferred from the estimated location of the user terminal and the calculated distance to the virtual anchor through a learned machine learning model; A location re-estimation unit that re-estimates the location of a user terminal based on the estimated location of the user terminal and the calculated metric values ​​of actual anchors and the inferred metric values ​​of virtual anchors; A positioning device using a radio map including a virtual anchor, which includes a .

7. In paragraph 6, A region-related anchor selection unit for selecting region-related anchors surrounding an area containing an estimated location of a user terminal; Including more, A positioning device using a radio map including a virtual anchor, wherein the distance calculation unit calculates a distance for real anchors and virtual anchors included in an area-related anchor, the real metric calculation unit calculates a metric value for real anchors included in an area-related anchor, the virtual metric inference unit infers a metric value for virtual anchors included in an area-related anchor, and the location re-estimation unit re-estimates the location of the user terminal based on the estimated location of the user terminal and the calculated metric values ​​of the real anchors included in the area-related anchor and the inferred metric values ​​of the virtual anchors included in the area-related anchor.

8. In paragraph 6, A positioning device using a radio map including a virtual anchor, wherein the machine learning model is trained to calculate a correlation coefficient or Euclidean distance between a measurement vector of a user terminal and a measurement vector of the corresponding anchor, using the estimated position of the user terminal and the calculated distance from the actual anchor as input data.

9. In paragraph 6, A positioning device using a radio map including virtual anchors, wherein the number of virtual anchors is determined based on the quality of service (QoS) of a positioning-based service provided by the positioning device.

10. In paragraph 9, A positioning device using a radio map including virtual anchors, wherein the number of virtual anchors is determined by referring to a lookup table storing a predetermined number of virtual anchors according to the quality of service (QoS) required for each positioning-based service provided by the positioning device.

11. In paragraph 6, A positioning device using a radio map including a virtual anchor, wherein the signal information constituting the measurement vector is one of positioning information including RSSI, RSRP, RSRQ, AOA, TOF, TOA, TDOA, CIR, and CFR.

12. In the method of constructing a radio map by a positioning device, A step of receiving measurement vectors from M real anchors; A step of storing position information and measurement vectors of M actual anchors; A step of storing location information of N virtual anchors that do not transmit signals; and A step of constructing a radio map including M real anchors and N virtual anchors; A method of configuring a radio map including a virtual anchor, comprising:

13. In paragraph 12, A method for constructing a radio map including virtual anchors, wherein the number of virtual anchors is determined based on the quality of service (QoS) of a positioning-based service provided by a positioning device.

14. In paragraph 13, A method for configuring a radio map including virtual anchors, wherein a positioning device determines the number of virtual anchors by referring to a lookup table in which a predetermined number of virtual anchors is stored according to the quality of service (QoS) required for each positioning-based service provided.

Images