Method and appartus for gru-based 3D localization using channel model

Abstract

A GRU-based three-dimensional spatial location estimation method using a channel model according to an embodiment of the present invention comprises: receiving an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space; inputting the ultra-wideband signal into a channel model classifier to determine one of a plurality of channel environments; and estimating the three-dimensional spatial location of the ultra-wideband signal transmitter based on a GRU-based transmitter location estimation model corresponding to the ultra-wideband signal and the determined channel environment.

Description

Technology Field

[0001] The present invention relates to a method and apparatus for estimating the three-dimensional position of an ultra-wideband signal receiver using a channel model and based on GRU. Background Technology

[0002] With the advancement of IoT (Internet of Things) technology, the demand for accurate indoor localization systems is increasing. Indoor localization systems can be applied to various applications, such as user tracking within buildings and the control of unmanned vehicles in warehouses.

[0003] However, in the case of indoor positioning systems based on deep learning models such as CNNs, it is necessary to perform training for every specific environment. That is, if the environment of the indoor positioning system changes, the deep learning model must be retrained with new data from that new environment.

[0004] Therefore, there is a growing need for a GRU-based three-dimensional spatial position estimation method and device that utilizes a channel model usable without the need for retraining in various environments. The problem to be solved

[0005] According to one embodiment of the present invention, a GRU-based three-dimensional spatial position estimation method and apparatus using a channel model that can be utilized without the need to perform retraining in various environments are provided. means of solving the problem

[0006] A GRU-based three-dimensional spatial location estimation method using a channel model according to an embodiment of the present invention comprises: receiving an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space; inputting the ultra-wideband signal into a channel model classifier to determine one of a plurality of channel environments; and estimating the three-dimensional spatial location of the ultra-wideband signal transmitter based on a GRU-based transmitter location estimation model corresponding to the ultra-wideband signal and the determined channel environment.

[0007] Preferably, between the step of receiving the ultra-wideband signal and the step of determining the channel environment, the method further includes a step of preprocessing the ultra-wideband signal to generate a GRU input value, and the step of determining the channel environment and the step of estimating the three-dimensional spatial position may use the GRU input value.

[0008] Preferably, the step of preprocessing the ultra-wideband signal to generate a GRU input value may include: generating a sampling signal by sampling the ultra-wideband signal at a predetermined frequency; normalizing the sampling signal to generate a real vector; downsampling the real vector to generate a downsampling signal; and combining the downsampling signals to generate the GRU input value.

[0009] Preferably, the step of determining the channel environment may utilize a GRU-based channel model classifier.

[0010] Preferably, the step of estimating the three-dimensional spatial position may involve selecting a GRU-based transmitter position estimation model corresponding to the determined channel environment among a plurality of GRU-based transmitter position estimation models, and inputting the ultra-wideband signal into the selected GRU-based transmitter position estimation model to estimate the three-dimensional spatial position.

[0011] In addition, a GRU-based three-dimensional spatial position estimation device using a channel model according to an embodiment of the present invention includes: a receiver unit that receives an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space; an environment determination unit that inputs the ultra-wideband signal into a channel model classifier to determine one of a plurality of channel environments; and a position estimation unit that estimates the three-dimensional spatial position of the ultra-wideband signal transmitter based on a GRU-based transmitter position estimation model corresponding to the ultra-wideband signal and the determined channel environment.

[0012] Preferably, the system further includes a preprocessing unit that preprocesses the ultra-wideband signal to generate a GRU input value, and the environment determination unit and the position estimation unit may use the GRU input value.

[0013] Preferably, the preprocessing unit may generate a sampling signal by sampling the ultra-wideband signal at a predetermined frequency, normalize the sampling signal to generate a real vector, downsample the real vector to generate a downsampling signal, and combine the downsampling signals to generate the GRU input value.

[0014] Preferably, the environment determination unit may use a GRU-based channel model classifier.

[0015] Preferably, the position estimation unit can select a GRU-based transmitter position estimation model corresponding to the determined channel environment among a plurality of GRU-based transmitter position estimation models, and input the ultra-wideband signal into the selected GRU-based transmitter position estimation model to estimate a three-dimensional spatial position. Effects of the invention

[0016] A GRU-based three-dimensional spatial position estimation method and device using a channel model according to one embodiment of the present invention have the effect of being immediately usable in various environments without the need for retraining. Brief explanation of the drawing

[0017] FIG. 1 is a flowchart illustrating a GRU-based three-dimensional spatial position estimation method using a channel model according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating a GRU-based three-dimensional spatial position estimation method using a channel model according to another embodiment of the present invention. FIG. 3 is a flowchart illustrating a method for generating GRU input values ​​according to an embodiment of the present invention. FIG. 4 is a block diagram illustrating a GRU-based three-dimensional spatial position estimation device using a channel model according to one embodiment of the present invention. FIG. 5 is a diagram illustrating an ultra-wideband signal transmitter and receiver according to an embodiment of the present invention. FIG. 6 is a diagram illustrating a channel model classifier according to an embodiment of the present invention. FIG. 7 is a diagram illustrating a GRU-based transmitter location estimation model according to one embodiment of the present invention. FIG. 8 is a diagram illustrating a GRU-based three-dimensional spatial position estimation device using a channel model according to an embodiment of the present invention. FIG. 9 is a diagram illustrating a channel model according to an embodiment of the present invention. Specific details for implementing the invention

[0018] The various embodiments described in this document are illustrative for the purpose of clearly explaining the technical concept of the present disclosure and are not intended to limit it to specific embodiments. The technical concept of the present disclosure includes various modifications, equivalents, alternatives, and embodiments optionally combined from all or part of each embodiment described in this document. Furthermore, the scope of the technical concept of the present disclosure is not limited to the various embodiments presented below or the specific descriptions thereof.

[0019] Terms used in this document, including technical or scientific terms, may have the meaning generally understood by those skilled in the art to which this disclosure pertains, unless otherwise defined.

[0020] Expressions used in this document, such as “includes,” “may include,” “is equipped,” “is equipped,” “has,” and “may have,” imply the existence of the subject feature (e.g., function, operation, or component, etc.) and do not exclude the existence of other additional features. In other words, such expressions should be understood as open-ended terms implying the possibility of including other embodiments.

[0021] Singular expressions used in this document may include the meaning of the plural form unless the context otherwise indicates, and this applies likewise to singular expressions described in the claims.

[0022] Expressions such as "first," "second," or "first," "second" used in this document are used to distinguish one object from another when referring to multiple objects of the same kind, unless the context implies otherwise, and do not limit the order or importance of said objects.

[0023] Expressions used in this document such as “A, B, and C,” “A, B, or C,” “at least one of A, B, and C,” or “at least one of A, B, or C” may refer to each of the listed items or all possible combinations of the listed items. For example, “at least one of A or B” may refer to (1) at least one A, (2) at least one B, and (3) at least one A and at least one B.

[0024] The expression "based on" as used in this document is used to describe one or more factors influencing an act or action of a decision or judgment described in the phrase or sentence containing such expression, and this expression does not exclude additional factors influencing said act or action of a decision or judgment.

[0025] Hereinafter, various embodiments of the present disclosure will be described with reference to the attached drawings. In the attached drawings and the description thereof, the same reference numerals may be assigned to identical or substantially equivalent components. Furthermore, in the description of the various embodiments below, the description of identical or corresponding components may be omitted, but this does not mean that such components are not included in the embodiments.

[0026] FIG. 1 is a flowchart illustrating a GRU-based three-dimensional spatial position estimation method using a channel model according to an embodiment of the present invention.

[0027] In step S110, the position estimation device receives an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space.

[0028] For example, referring to FIG. 5, a closed space with width D1, depth D2, and height D3 is shown. Within that space, a transmitter (Tx) that transmits an ultra-wideband (UWB) signal is located, and eight receivers (Rx) located at eight corners can receive the ultra-wideband signal.

[0029] In step S120, the location estimation device inputs the ultra-wideband signal into a channel model classifier to determine one of the multiple channel environments.

[0030] At this time, the channel model classifier may be a deep learning model. By inputting the ultra-wideband signal into the deep learning model, the location estimation device can classify which channel environment the ultra-wideband signal belongs to among multiple channel environments.

[0031] In another embodiment, the location estimation device may use a GRU-based channel model classifier.

[0032] For example, referring to FIG. 6, the channel model classifier may consist of an input side, one GRU layer, two fully connected layers, a softmax layer, and a classification layer. The GRU cell of the GRU layer uses the result of the GRU cell obtained in the previous time step along with the current input value. The length of the GRU cell result is the number of units in the GRU layer, and the number of units in the GRU layer may be 32. Additionally, the last result of the GRU layer is used as the input to the fully connected layer, and the number of units in the two fully connected layers may be 128 and 9. Subsequently, the classification degree for a total of 9 channel environments is determined in the softmax layer, and the channel environment corresponding to the largest value among the softmax layer results can be finally selected in the classification layer.

[0033] Meanwhile, referring to FIG. 9, in one embodiment of the present invention, nine channel models can be defined for a total of nine channel environments. Here, LoS represents the line of sight path, and NLoS represents the non-line of sight path. Furthermore, the channel models according to one embodiment of the present invention can be modeled as residential, office, outdoor, and industrial environments modeled with LoS and NLoS.

[0034] Finally, in step S130, the position estimation device estimates the three-dimensional spatial position of the ultra-wideband signal transmitter based on a GRU-based transmitter position estimation model corresponding to the ultra-wideband signal and the determined channel environment.

[0035] At this time, the position estimation device may have multiple GRU-based transmitter position estimation models that correspond to each of the multiple channel environments and are pre-trained according to the individual channel environment. Then, the position estimation device can estimate the three-dimensional spatial position of the ultra-wideband signal transmitter by using the ultra-wideband signal and the estimation model corresponding to the determined channel environment among the multiple GRU-based transmitter position estimation models.

[0036] For example, referring to FIG. 7, the GRU-based transmitter position estimation model may consist of an input layer, two GRU layers composed of GRU cells having 64 and 32 units, three fully connected layers having 256, 32, and 3 units, and a regression layer. The result of the first GRU layer can be used as the input to the second GRU layer. The result of the second GRU layer is used as the input to the first fully connected layer, and finally, the position of the transmitter transmitting the ultra-wideband signal is estimated in the regression layer. At this time, the loss function of the GRU-based transmitter position estimation model can be defined as Equation 1.

[0037] [Mathematical Formula 1]

[0038]

[0039] Here, is the ground truth, and is the estimate of the transmitter (Tx) position for the q-th training sample of the transmitter. The training process of the GRU-based transmitter position estimation model is the loss function It is an attempt to minimize.

[0040] In another embodiment, when a position estimation device estimates a three-dimensional spatial position, it may select a GRU-based transmitter position estimation model corresponding to the determined channel environment among a plurality of GRU-based transmitter position estimation models, and input the ultra-wideband signal into the selected GRU-based transmitter position estimation model to estimate the three-dimensional spatial position.

[0041] For example, the position estimation device can select a specific GRU-based transmitter position estimation model corresponding to the determined channel environment from among nine GRU-based transmitter position estimation models corresponding to each of the nine channel models. Then, by inputting the ultra-wideband signal into the corresponding specific GRU-based transmitter position estimation model, the position estimation device can estimate the three-dimensional spatial location of the transmitter that transmitted the ultra-wideband signal.

[0042] More specifically, referring to FIG. 8, the position estimation device can determine the channel environment by inputting the ultra-wideband signal (Uo) into a GRU-based channel model classifier. Then, the position estimation device can estimate the three-dimensional spatial position of the transmitter (Tx) by selecting a GRU-based transmitter position estimation model corresponding to the determined channel environment from among GRU-based transmitter position estimation models from CM-1-NET to CM-9-NET, and inputting the ultra-wideband signal (Uo) into the selected GRU-based transmitter position estimation model.

[0043] FIG. 2 is a flowchart illustrating a GRU-based three-dimensional spatial position estimation method using a channel model according to another embodiment of the present invention.

[0044] In step S210, the position estimation device receives an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space.

[0045] In step S220, the position estimation device preprocesses the ultra-wideband signal to generate a GRU input value.

[0046] At this time, the preprocessing performed by the position estimation device may be for the purpose of making the ultra-wideband signal suitable for use in a GRU-based channel model classifier and a GRU-based transmitter position estimation model. In addition, the specific process will be described in detail later in the description of Fig. 3.

[0047] In step S230, the position estimation device inputs its GRU input value into a channel model classifier to determine one of the multiple channel environments.

[0048] Finally, in step S240, the position estimation device estimates the three-dimensional spatial position of the ultra-wideband signal transmitter based on the GRU input value and the GRU-based transmitter position estimation model corresponding to the determined channel environment.

[0049] FIG. 3 is a flowchart illustrating a method for generating GRU input values ​​according to an embodiment of the present invention.

[0050] In step S310, the position estimation device generates a sampling signal by sampling an ultra-wideband signal at a predetermined frequency.

[0051] For example, a position estimation device uses the p-th receiver in a channel environment corresponding to the o-th channel model to obtain the q-th sample, which is an ultra-wideband signal ( It can receive ). At this time, And, And, It can be. The position estimation device uses an ultra-wideband signal ( ) sampling frequency f s Sampling the sampling signal ( It can generate ). At this time, And, It can be. (e.g., N=3600)

[0052] In step S320, the position estimation device normalizes the sampling signal to generate a real vector.

[0053] That is, the position estimation device is a sampling signal ( Normalize ) to create a real vector ( It can generate ). In this case, the real vector can be defined as Equation 2 below.

[0054] [Mathematical Formula 2]

[0055]

[0056] Here, and are each It is the mean and standard deviation of.

[0057] In step S330, the position estimation device downsamples the real vector to generate a downsampled signal.

[0058] At this time, the position estimation device uses the real vector ( ) to r d Downsampling the downsampled signal ( Can generate ).

[0059] Meanwhile, time step n r 째's The value can be expressed as shown in mathematical formula 3 below.

[0060] [Mathematical Formula 3]

[0061]

[0062] Here, am.

[0063] In step S340, the position estimation device combines the downsampled signals to generate a GRU input value.

[0064] For example, a position estimation device can generate a GRU input value by combining all downsampled signals. In this case, the GRU input value can be expressed as Equation 4.

[0065] [Mathematical Formula 4]

[0066]

[0067] Here, It can be divided into multiple column vectors according to the time step.

[0068] FIG. 4 is a block diagram illustrating a GRU-based three-dimensional spatial position estimation device using a channel model according to an embodiment of the present invention.

[0069] Referring to FIG. 4, a GRU-based three-dimensional spatial position estimation device (400) using a channel model according to one embodiment of the present invention includes a receiving unit (410), a preprocessing unit (420), an environment determination unit (430), and a position estimation unit (440).

[0070] The receiver (410) receives an ultra-wideband signal from an ultra-wideband signal transmitter using multiple receivers in a three-dimensional indoor space.

[0071] The preprocessing unit (420) preprocesses the ultra-wideband signal to generate a GRU input value.

[0072] The environment determination unit (430) inputs the GRU input value into a channel model classifier to determine one of the multiple channel environments.

[0073] Finally, the position estimation unit (440) estimates the three-dimensional spatial position of the ultra-wideband signal transmitter based on the GRU input value and the GRU-based transmitter position estimation model corresponding to the determined channel environment.

[0074] In another embodiment, the preprocessing unit (420) may generate a sampling signal by sampling an ultra-wideband signal at a predetermined frequency, normalize the sampling signal to generate a real vector, downsample the real vector to generate a downsampling signal, and combine the downsampling signals to generate a GRU input value.

[0075] In another embodiment, the environment determination unit (430) may use a GRU-based channel model classifier.

[0076] In another embodiment, the position estimation unit (440) can select a GRU-based transmitter position estimation model corresponding to the determined channel environment among a plurality of GRU-based transmitter position estimation models, and input the GRU input value into the selected GRU-based transmitter position estimation model to estimate a three-dimensional spatial position.

[0077] Although the technical concept according to the disclosure of this specification has been described by various embodiments above, the technical concept according to the disclosure of this specification includes various substitutions, modifications, and changes that can be understood by a person skilled in the art to which the disclosure of this specification belongs. Furthermore, it should be understood that such substitutions, modifications, and changes may be included within the scope of the appended claims.

Claims

Claim 1 A GRU-based three-dimensional spatial position estimation method using a channel model, comprising: a step of receiving an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space; a step of inputting the ultra-wideband signal into a channel model classifier to determine one channel environment among a plurality of channel environments; and a step of estimating the three-dimensional spatial position of the ultra-wideband signal transmitter based on the ultra-wideband signal and a GRU-based transmitter position estimation model corresponding to the determined channel environment, wherein different GRU-based transmitter position estimation models are defined for each preset channel environment, and transmitter position estimation models for the preset channel environments are pre-trained and generated to match individual channel environments, and one transmitter position estimation model corresponding to the determined channel environment is selected among the transmitter position estimation models, and the ultra-wideband signal is input into the selected transmitter position estimation model to estimate the three-dimensional spatial position of the transmitter. Claim 2 A GRU-based three-dimensional spatial position estimation method using a channel model according to claim 1, wherein, between the step of receiving the ultra-wideband signal and the step of determining the channel environment, the method further includes the step of preprocessing the ultra-wideband signal to generate a GRU input value, and the step of determining the channel environment and the step of estimating the three-dimensional spatial position utilize the GRU input value. Claim 3 A GRU-based three-dimensional spatial position estimation method using a channel model according to claim 2, wherein the step of preprocessing the ultra-wideband signal to generate a GRU input value comprises: generating a sampling signal by sampling the ultra-wideband signal at a predetermined frequency; normalizing the sampling signal to generate a real vector; downsampling the real vector to generate a downsampling signal; and combining the downsampling signals to generate the GRU input value. Claim 4 A GRU-based three-dimensional spatial location estimation method using a channel model, wherein the step of determining the channel environment in claim 1 is characterized by using a GRU-based channel model classifier. Claim 5 delete Claim 6 A GRU-based three-dimensional spatial position estimation device using a channel model, comprising: a receiver unit that receives an ultra-wideband signal from an ultra-wideband signal transmitter using a plurality of receivers in a three-dimensional indoor space; an environment determination unit that inputs the ultra-wideband signal into a channel model classifier to determine one channel environment among a plurality of channel environments; and a position estimation unit that estimates the three-dimensional spatial position of the ultra-wideband signal transmitter based on the ultra-wideband signal and a GRU-based transmitter position estimation model corresponding to the determined channel environment, wherein different GRU-based transmitter position estimation models are defined for each preset channel environment, and transmitter position estimation models for the preset channel environments are pre-learned and generated to match individual channel environments, and the position estimation unit selects one transmitter position estimation model corresponding to the determined channel environment among the transmitter position estimation models, and inputs the ultra-wideband signal into the selected transmitter position estimation model to estimate the three-dimensional spatial position of the transmitter. Claim 7 A GRU-based three-dimensional spatial position estimation device using a channel model, characterized in that, in claim 6, it further includes a preprocessing unit that preprocesses the ultra-wideband signal to generate a GRU input value, and the environment determination unit and the position estimation unit utilize the GRU input value. Claim 8 A GRU-based three-dimensional spatial position estimation device using a channel model, characterized in that, in claim 7, the preprocessing unit generates a sampling signal by sampling the ultra-wideband signal at a predetermined frequency, normalizes the sampling signal to generate a real vector, downsamples the real vector to generate a downsampling signal, and combines the downsampling signals to generate the GRU input value. Claim 9 In claim 6, the above-mentioned environment determination unit is characterized by using a GRU-based channel model classifier, and the GRU-based three-dimensional spatial position estimation device using a channel model is characterized by the above-mentioned environment determination unit using a GRU-based channel model classifier. Claim 10 delete

Images