Method and apparatus for augmenting data of UWB signal on basis of artificial intelligence model

Abstract

The present disclosure relates to a method for augmenting data of UWB signals on the basis of an artificial intelligence model. The method for augmenting data of UWB signals on the basis of an artificial intelligence model comprises the steps of: receiving a first UWB signal and a second UWB signal; generating a third UWB signal by performing interpolation on the basis of the first UWB signal and the second UWB signal; and providing at least one of the first UWB S signal, the second UWB S signal, and the third UWB S signal to a trained artificial intelligence model, thereby generating a fourth UWB signal having a distribution associated with the at least one signal.

Description

Method and device for data augmentation of UWB signals based on artificial intelligence models

[0001] The present disclosure relates to a method and device for data augmentation of a UWB signal based on an artificial intelligence model, and more particularly, to a method and device for augmenting data representing a UWB signal, which is two-dimensional data of time and amplitude.

[0002] Recently, artificial intelligence technology has been utilized in various industrial fields, including object recognition. As AI technology advances, securing high-quality training data for training AI models is becoming increasingly important. Accordingly, data augmentation techniques are being used to secure high-quality training data.

[0003] Data augmentation techniques apply artificial transformations to original data to acquire large amounts of new data. However, because UWB signals are two-dimensional, they contain significantly less information than three-dimensional data. Therefore, generating data using typical data augmentation techniques can result in data that is nearly identical to the original data. In such cases, when there is no difference between the original and augmented data, there is a high risk of overfitting of the AI ​​model during the learning process.

[0004] The present invention is derived from research conducted as part of the ICT Convergence Industry Innovation Technology Development Project of the Ministry of Science and ICT (Project Unique Number: 1711193657, Project Number: 2017-0-00418-007, Project Management Agency: Information and Communications Technology Planning and Evaluation Institute, Research Project Title: Research on Time-Domain Artificial Intelligence Radar SoC (System On a Chip) Design Using Ultra-High-Speed ​​Sampling Technique, Project Performing Agency: Yonsei University Industry-Academic Cooperation Foundation, Research Period: 2023.01.01 - 2023.12.31).

[0005] Meanwhile, the Korean government, which provided the task, has no property interest in any aspect of the present invention.

[0006] The present disclosure provides a data augmentation method of a UWB signal based on an artificial intelligence model to solve the above-described problems, a computer program stored in a computer-readable medium, a computer-readable medium storing the computer program, and a device (system).

[0007] In various embodiments of the present disclosure, when using UWB signals in the time domain, features can be effectively extracted not only for dynamically moving objects but also for stationary objects.

[0008] In various embodiments of the present disclosure, a computing device can effectively generate high-quality data required for model learning by first augmenting data through an interpolation algorithm such as SMOTE, and secondarily augmenting data through a conditional adversarial generative neural network, thereby significantly reducing the costs required for data collection and model learning.

[0009] In various embodiments of the present disclosure, a computing device can generate a large amount of training data necessary for training an artificial intelligence model for object recognition by amplifying a small amount of UWB signals measured using a UWB device.

[0010] Embodiments of the present disclosure will be described below with reference to the accompanying drawings, wherein like reference numerals represent similar elements, but are not limited thereto.

[0011] FIG. 1 is a diagram illustrating an example of generating a UWB signal associated with an object according to one embodiment of the present disclosure.

[0012] FIG. 2 is a functional block diagram showing the internal configuration of a computing device according to one embodiment of the present disclosure.

[0013] FIG. 3 is an exemplary block diagram showing a process in which data augmentation is performed according to one embodiment of the present disclosure.

[0014] FIG. 4 is an exemplary diagram showing an artificial neural network according to one embodiment of the present disclosure.

[0015] FIG. 5 is an exemplary block diagram showing the detailed structure of an artificial intelligence model according to one embodiment of the present disclosure.

[0016] FIG. 6 is a flowchart illustrating an example of a data augmentation method of a UWB signal based on an artificial intelligence model according to one embodiment of the present disclosure.

[0017] FIG. 7 is a block diagram showing a hardware configuration of a computing device according to one embodiment of the present disclosure.

[0018] Figure 8 is a drawing for explaining the effect according to the present disclosure.

[0019] The present disclosure can be implemented in various ways, including a method, a device (system), a computer program stored on a computer-readable medium, or a computer-readable medium having a computer program stored thereon.

[0020] According to one embodiment of the present disclosure, a method for data augmentation of a UWB signal based on an artificial intelligence model performed by at least one processor includes the steps of receiving a first UWB signal and a second UWB signal, generating a third UWB signal based on the first UWB signal and the second UWB signal using an interpolation algorithm, and providing at least one signal among the first UWB signal, the second UWB signal, and the third UWB signal to a learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal.

[0021] According to one embodiment of the present disclosure, the step of generating a third UWB signal includes the step of generating a third UWB signal based on third label data located between first label data corresponding to a first UWB signal and second label data corresponding to a second UWB signal in a data space.

[0022] According to one embodiment of the present disclosure, the step of generating the third UWB signal includes the step of generating the third UWB signal using a SMOTE algorithm.

[0023] According to one embodiment of the present disclosure, the artificial intelligence model is a model based on a conditional adversarial generative neural network including a generator and a discriminator.

[0024] According to one embodiment of the present disclosure, the generator has a 1D U-Net structure.

[0025] According to one embodiment of the present disclosure, the generator is trained to generate a fake UWB signal having a distribution similar to at least one of a first UWB signal, a second UWB signal, and a third UWB signal in response to receiving a sampling vector and a condition vector corresponding to a class of UWB signals. Furthermore, the discriminator is trained to determine whether the fake UWB signal is genuine or not in response to receiving the generated fake UWB signal.

[0026] According to one embodiment of the present disclosure, the discriminator includes an auxiliary classifier for classifying a class of fake UWB signals.

[0027] According to one embodiment of the present disclosure, the first UWB signal and the second UWB signal are generated by measuring an object at the same angle.

[0028] A computer program stored in a computer-readable recording medium is provided for executing the above-described method according to one embodiment of the present disclosure on a computer.

[0029] A computing device according to one embodiment of the present disclosure includes a communication module, a memory, and at least one processor connected to the memory and configured to execute at least one computer-readable program contained in the memory. The at least one program includes instructions for receiving a first UWB signal and a second UWB signal, generating a third UWB signal based on the first UWB signal and the second UWB signal using an interpolation algorithm, and providing at least one of the first UWB signal, the second UWB signal, and the third UWB signal to a learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal.

[0030] Hereinafter, specific details for implementing the present disclosure will be described in detail with reference to the attached drawings. However, in the following description, specific descriptions of widely known functions or configurations will be omitted if they may unnecessarily obscure the gist of the present disclosure.

[0031] In the attached drawings, identical or corresponding components are assigned the same reference numerals. Furthermore, in the description of the embodiments below, duplicate descriptions of identical or corresponding components may be omitted. However, even if a description of a component is omitted, it is not intended that such component is not included in any embodiment.

[0032] The advantages and features of the disclosed embodiments, and the methods for achieving them, will become clearer with reference to the embodiments described below, along with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided solely to ensure the completeness of the present invention and to fully inform those skilled in the art of the scope of the invention.

[0033] The terms used in this specification will be briefly explained, followed by a detailed description of the disclosed embodiments. The terms used in this specification have been selected from widely used, current terms, taking into account the functions of the present invention. However, these terms may vary depending on the intentions of engineers working in the relevant fields, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this specification should not be defined simply based on their names, but rather based on their meanings and the overall content of the present disclosure.

[0034] In this specification, singular expressions include plural expressions unless the context clearly indicates otherwise. Furthermore, plural expressions include singular expressions unless the context clearly indicates otherwise. When a part of the specification is said to include a component, this does not exclude other components, but rather implies that other components may be included, unless otherwise specifically stated.

[0035] In this disclosure, terms such as “comprise,” “comprising,” and the like may indicate the presence of features, steps, operations, elements, and / or components, but such terms do not exclude the addition of one or more other features, steps, operations, elements, components, and / or combinations thereof.

[0036] In this disclosure, when a particular component is referred to as being "coupled," "combined," "connected," or "reacting" with any other component, the particular component may be directly coupled, combined, and / or connected to, or reacting with, the other component, but is not limited thereto. For example, one or more intermediate components may exist between the particular component and the other component. Furthermore, "and / or" in the present disclosure may include each of one or more of the listed items, or a combination of at least some of one or more of the listed items.

[0037] In this disclosure, terms such as "first," "second," etc. are used to distinguish specific components from other components, and the components described by these terms are not limited. For example, the "first" component may be an element of the same or similar form as the "second" component.

[0038] In the present disclosure, "UWB (ultra-wideband) signal" refers to an ultra-wideband signal, and may refer to two-dimensional data in the time domain composed of time and amplitude. Here, the UWB signal is not limited to referring to a single piece of data, and may include multiple pieces of data. In addition, "UWB device" may refer to a radar device that generates or receives a UWB signal.

[0039] In the present disclosure, "SMOTE (synthetic minority over-sampling technique) or SMOTE algorithm" may refer to one of the interpolation and / or over-sampling algorithms, which is intended to prevent the problem of duplicate generation of completely identical data. SMOTE may perform interpolation by selecting the k closest neighboring vectors to a specific vector among data of an arbitrary class, connecting a reference vector and the selected vectors with a line segment, and extracting an arbitrary point on the line segment as a new vector.

[0040] In the present disclosure, “conditional generative adversarial nets (CGAN)” may refer to an artificial intelligence model that adds conditional information to a general generative adversarial neural network to limit the classes of data to be output.

[0041]

[0042] FIG. 1 is a diagram illustrating an example of generating a UWB signal associated with an object (100) according to one embodiment of the present disclosure. According to one embodiment, the UWB signal may be used to determine the location of the object (100) and extract characteristics of the object (100). Here, the UWB signal may refer to an impulse signal that is emitted from any UWB device, hits the object (100), and returns. For example, each UWB device may calculate the distance to the object (100) by utilizing information on the transmission time and the reception time, or may extract characteristics of the object (100) by utilizing the Doppler effect.

[0043] The illustrated example may illustrate an example in which each UWB device detects or measures an object (100) at positions of 0 degrees, 90 degrees, 180 degrees, and 270 degrees, respectively. For example, a first UWB device (102) may generate a UWB signal by emitting a pulse at a first angle corresponding to 0 degrees, a second UWB device (104) may generate a UWB signal by emitting a pulse at a second angle corresponding to 90 degrees, and a third UWB device (106) may generate a UWB signal by emitting a pulse at a third angle corresponding to 180 degrees, and a fourth UWB device (108) may generate a UWB signal by emitting a pulse at a fourth angle corresponding to 270 degrees. When UWB signals are acquired at various angles in this way, features of the object (100) at various angles and / or sides may all be detected or measured.

[0044] According to one embodiment, the UWB signal may be two-dimensional data in the time domain consisting of time and amplitude. Here, the time domain may refer to a domain for focusing on the temporal characteristics of the waveform, consisting of the x-axis representing the passage of time and the y-axis representing the amplitude. In this way, when using the UWB signal in the time domain, the characteristics of not only dynamically moving objects but also stationary objects (100) can be effectively extracted.

[0045]

[0046] FIG. 2 is a functional block diagram showing the internal configuration of a computing device (200) according to one embodiment of the present disclosure. According to one embodiment, the computing device (200) is any device for performing data augmentation of a UWB signal, and may include a signal receiving unit (202), an interpolation performing unit (204), a model learning unit (206), a data augmentation unit (208), etc. For example, the computing device (200) may perform data augmentation of a UWB signal to secure a large amount of learning data (e.g., UWB signal data) for training an arbitrary artificial intelligence model for object recognition.

[0047] According to one embodiment, the signal receiving unit (202) can receive a UWB signal that has measured an arbitrary object. For example, when an arbitrary UWB device emits a pulse signal to an object and then measures a UWB signal that returns, the signal receiving unit (202) can receive the UWB signal measured in this way. Here, the UWB signal can include multiple signals measured at one position and angle, or multiple signals measured at multiple positions and angles. For example, the signal receiving unit (202) can receive a first UWB signal and a second UWB signal that have repeatedly measured an object at the same angle.

[0048] According to one embodiment, the computing device (200) and / or the signal receiving unit (202) may perform preprocessing on the received UWB signal. For example, the computing device (200) and / or the signal receiving unit (202) may perform labeling on the UWB signal based on the angle at which the UWB signal is measured, thereby converting the label data (e.g., data in vector form) in the data space. In this case, UWB signals measured at the same angle may be assigned the same class to form a distribution in the data space.

[0049] According to one embodiment, the interpolation unit (204) may augment the UWB signal and / or label data associated with the UWB signal using an interpolation algorithm. Here, the interpolation algorithm may refer to a method or algorithm for supplementing or estimating the value of missing or undetermined data using known data. For example, the interpolation unit (204) may generate a third UWB signal based on the first UWB signal and the second UWB signal using the interpolation algorithm. Here, the third UWB signal may represent a signal having the same class as the first UWB signal and the second UWB signal.

[0050] According to one embodiment, the interpolation unit (204) may generate a third UWB signal based on third label data located between first label data corresponding to the first UWB signal and second label data corresponding to the second UWB signal in the data space. For example, the interpolation unit (204) may generate the third UWB signal using a synthetic minority over-sampling technique (SMOTE) algorithm. Here, SMOTE or the SMOTE algorithm is one of interpolation and / or over-sampling algorithms, and is an algorithm for preventing the problem of duplicate generation of completely identical data, and may refer to an algorithm that generates a value between two pieces of data as new data. That is, when the SMOTE algorithm is used, the interpolation unit (204) may generate the third UWB signal based on any point existing between a line segment connecting the first label data corresponding to the first UWB signal and the second label data corresponding to the second UWB signal.

[0051] According to one embodiment, the model learning unit (206) can perform learning of an artificial intelligence model for data augmentation using both the UWB signal received by the signal receiving unit (202) and the UWB signal generated by the interpolation performing unit (204). For example, the model learning unit (206) can provide the first UWB signal, the second UWB signal, and the third UWB signal to the artificial intelligence model for data augmentation to train the corresponding artificial intelligence model. That is, the model learning unit (206) can train the artificial intelligence model to generate a fourth UWB signal similar to the first UWB signal, the second UWB signal, and the third UWB signal.

[0052] In one embodiment, the AI ​​model may be a conditional adversarial generative neural network-based model including a generator and a discriminator. For example, the generator may be trained to generate fake UWB signals similar to input UWB signals, and the discriminator may be trained to determine whether the fake UWB signals generated by the generator are authentic. In other words, the generator and discriminator may be trained to generate data that is similar to, but different from, actual UWB signals by performing adversarial training with each other.

[0053] According to one embodiment, the artificial intelligence model can be trained based on a loss function as in the following mathematical expression 1.

[0054]

[0055]

[0056]

[0057] Here, x is the actual data Image sampled from, z is a probability distribution A vector sampled from, E represents the expected value, D represents the discriminator, and G represents the generator. According to this mathematical expression 1, unlike the existing adversarial generative neural network, the artificial intelligence model can control the generated data by adding the constraint y to the loss term.

[0058] According to one embodiment, the data augmentation unit (208) may provide at least one of the first UWB signal, the second UWB signal, and the third UWB signal to the learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal. The data augmentation unit (208) may provide UWB signals measured by the UWB device or generated by an interpolation algorithm to the artificial intelligence model to generate a UWB signal similar to the input UWB signal. The UWB signals augmented in this way may be utilized as training data for an arbitrary model for recognizing the characteristics of an object.

[0059] In Fig. 2, each functional configuration included in the computing device (200) is separately described, but this is only to help understand the invention, and two or more functions may be performed in a single computing device. By this configuration, the computing device (200) can effectively generate high-quality data required for model learning by first augmenting data through an interpolation algorithm such as SMOTE, and secondarily augmenting data through a conditional adversarial generative neural network, and accordingly, the cost required for data collection and model learning can be significantly reduced.

[0060]

[0061] FIG. 3 is an exemplary block diagram illustrating a process of performing data augmentation according to one embodiment of the present disclosure. As described above, the interpolation performing unit (204) may perform data interpolation based on the first UWB signal (302) and the second UWB signal (304) to generate a third UWB signal (306). In addition, the data augmentation unit (208) may use a learned artificial intelligence model to generate a fourth UWB signal (308) associated with at least one of the first UWB signal (302), the second UWB signal (304), and the third UWB signal (306).

[0062] According to one embodiment, the first UWB signal (302) and the second UWB signal (304) may refer to different signals measured at the same position and angle, and may be expressed as label data in the data space. When the first UWB signal (302) and the second UWB signal (304) in vector form are received in this way, the interpolation performing unit (204) may generate the third UWB signal (306) based on any point located between the line segments connecting the first label data corresponding to the first UWB signal (302) and the second label data corresponding to the second UWB signal (304). When interpolation is performed in this way, the third UWB signal (306) may be generated, which includes characteristics that are different from, but similar to, the first UWB signal (302) and the second UWB signal (304). In FIG. 3, it is described that one third UWB signal (306) is generated based on the first UWB signal (302) and the second UWB signal (304), but the number of UWB signals provided to or generated by the interpolation performing unit (204) may be determined differently.

[0063] According to one embodiment, the data augmentation unit (208) can provide the first UWB signal (302), the second UWB signal (304), and the third UWB signal (306) to the learned artificial intelligence model to generate a fourth UWB signal (308). Here, the fourth UWB signal (308) may include characteristics that are different from, but similar to, the first UWB signal (302), the second UWB signal (304), and the third UWB signal (306). Although FIG. 3 describes that one fourth UWB signal (308) is generated, the number of UWB signals provided to or generated by the data augmentation unit (208) may be determined differently. With this configuration, the computing device (200 of FIG. 2) can generate a large amount of training data necessary for training an artificial intelligence model for object recognition by augmenting a small amount of UWB signals measured using the UWB device.

[0064]

[0065] FIG. 4 is an exemplary diagram illustrating an artificial neural network (400) according to one embodiment of the present disclosure. The artificial neural network (400) is an example of the artificial intelligence model described above, and in machine learning technology and cognitive science, is a statistical learning algorithm implemented based on the structure of a biological neural network or a structure that executes the algorithm.

[0066] According to one embodiment, the artificial neural network (400) may represent a machine learning model having problem-solving capabilities by learning that nodes, which are artificial neurons that form a network by combining synapses like in a biological neural network, repeatedly adjust the weights of synapses so that the error between the correct output corresponding to a specific input and the inferred output is reduced. For example, the artificial neural network (400) may include any language model used in artificial intelligence learning methods such as machine learning and deep learning.

[0067] According to one embodiment, the artificial neural network (400) can be implemented as a multilayer perceptron (MLP) composed of multiple layers of nodes and connections therebetween. The artificial neural network (400) according to the present embodiment can be implemented using one of the structures of various artificial intelligence models including MLP. As illustrated in FIG. 4, the artificial neural network (400) is composed of an input layer (420) that receives an input signal or data (410) from the outside, an output layer (440) that outputs an output signal or data (450) corresponding to the input data, and n hidden layers (430_1 to 430_n) located between the input layer (420) and the output layer (440) that receive signals from the input layer (420), extract characteristics, and transmit them to the output layer (440) (where, n is a positive integer). Here, the output layer (440) receives signals from the hidden layers (430_1 to 430_n) and outputs them to the outside.

[0068] The learning method of the artificial neural network (400) includes a supervised learning method that learns to optimize problem solving through input of a teacher signal (correct answer), and an unsupervised learning method that does not require a teacher signal. According to one embodiment, the artificial neural network (400) can be trained to generate an additional UWB signal by inputting an actual UWB signal measured by a UWB device and a UWB signal supplemented by an interpolation algorithm.

[0069]

[0070] FIG. 5 is an exemplary block diagram illustrating a detailed structure of an artificial intelligence model according to an embodiment of the present disclosure. As described above, the artificial intelligence model may be a conditional adversarial generative neural network-based model including a generator (510) and a discriminator (520). Here, the generator (510) may have a 1D U-Net structure. That is, the generator of the conditional adversarial generative neural network-based model may have a structure in which the number of channels increases and then decreases, and may be configured as a 1D convolution. In this way, when the generator (510) with a 1D U-Net structure is used, detailed features of the UWB signal can be precisely captured through multiple convolution layers, and gradient loss and overfitting can be prevented through skip connections.

[0071] According to one embodiment, the generator (510) may receive a sampling vector (e.g., a 500-dimensional vector) sampled from a normal distribution and a condition vector corresponding to a class of UWB signals (e.g., four classes corresponding to 0 degrees, 90 degrees, 180 degrees, and 270 degrees) to generate a fake UWB signal (512). Here, the fake UWB signal (512) is data generated during the learning process of the generator (510) and the discriminator (520), and may be fake data generated by the generator (510) similar to the input UWB signal.

[0072] According to one embodiment, the discriminator (520) can receive a fake UWB signal (512) generated by the generator (510) and determine whether the input fake UWB signal (512) is authentic. For example, the discriminator (520) can compare the fake UWB signal (512) with a correct UWB signal (514) to determine whether the fake UWB signal (512) is authentic. If the discriminator (520) determines that the fake UWB signal (512) is fake data, the generator (510) can generate a UWB signal more precisely so that the discriminator (520) can mistake the fake data for real data, and the discriminator (520) can be trained to more precisely determine whether the UWB signal generated by the generator (510) is authentic.

[0073] According to one embodiment, the discriminator (520) may further include an auxiliary classifier (530) for classifying the class of the fake UWB signal (512). For example, the auxiliary classifier (530) may classify the class of the fake UWB signal (512) generated by the generator (510) into a class corresponding to 0 degrees, 90 degrees, 180 degrees, or 270 degrees. When the auxiliary classifier (530) is additionally used in this way, the performance of the discriminator (520) in determining whether a signal is genuine or fake may be improved, and accordingly, the performance of the generator (510) in generating a precise UWB signal may also be further improved.

[0074] Although FIG. 5 describes four classes of UWB signals corresponding to 0 degrees, 90 degrees, 180 degrees, and 270 degrees, the present invention is not limited thereto, and the classes and number of classes of UWB signals may be determined differently depending on the position at which the UWB device measures an object. With this configuration, the artificial intelligence model can effectively generate a large number of UWB signals having various classes based on the generator (510) and the discriminator (520).

[0075]

[0076] FIG. 6 is a flowchart illustrating an example of a method (600) for data augmentation of a UWB signal based on an artificial intelligence model according to an embodiment of the present disclosure. The method (600) for data augmentation of a UWB signal based on an artificial intelligence model may be performed by a processor (e.g., at least one processor of a computing device). The method (600) for data augmentation of a UWB signal based on an artificial intelligence model may be initiated when the processor receives a first UWB signal and a second UWB signal (S610). Here, the first UWB signal and the second UWB signal may be generated by measuring an object at the same angle.

[0077] According to one embodiment, the processor may generate a third UWB signal based on the first UWB signal and the second UWB signal using an interpolation algorithm (S620). For example, the processor may generate the third UWB signal based on third label data between first label data corresponding to the first UWB signal and second label data corresponding to the second UWB signal in the data space. In this case, the processor may generate the third UWB signal using the SMOTE algorithm.

[0078] According to one embodiment, the processor may provide at least one of the first UWB signal, the second UWB signal, and the third UWB signal to the learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal (S630). Here, the data space may refer to a space including label data generated by performing labeling on the UWB signal, and the distribution associated with the signal may refer to a distribution formed by gathering label data having the same class in the data space. That is, label data having the same class may form a distribution.

[0079] According to one embodiment, the artificial intelligence model may be a conditional adversarial generative neural network-based model including a generator and a discriminator. Here, the generator may have a 1D U-Net structure. In this case, the generator may be trained to generate a fake UWB signal having a distribution similar to at least one of a first UWB signal, a second UWB signal, and a third UWB signal in response to receiving a sampling vector and a condition vector corresponding to a class of a UWB signal. In addition, the discriminator may be trained to determine whether the fake UWB signal is genuine in response to receiving the generated fake UWB signal. Additionally, the discriminator may include an auxiliary classifier for classifying a class of the fake UWB signal, and the auxiliary classifier may be trained to classify the class of the UWB signal.

[0080]

[0081] FIG. 7 is a block diagram illustrating a hardware configuration of a computing device (200) according to one embodiment of the present disclosure. The computing device (200) may include a memory (710), a processor (720), a communication module (730), and an input / output interface (740). As illustrated in FIG. 7, the computing device (200) may be configured to communicate information and / or data via a network using the communication module (730).

[0082] The memory (710) may include any non-transitory computer-readable recording medium. In one embodiment, the memory (710) may include a non-volatile mass storage device such as a random access memory (RAM), a read only memory (ROM), a disk drive, a solid state drive (SSD), a flash memory, etc. As another example, a non-volatile mass storage device such as a ROM, an SSD, a flash memory, a disk drive, etc. may be included in the computing device (200) as a separate permanent storage device distinct from the memory. In addition, an operating system and at least one program code may be stored in the memory (710).

[0083] These software components may be loaded from a computer-readable recording medium separate from the memory (710). This separate computer-readable recording medium may include a recording medium directly connectable to the computing device (200), for example, a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD / CD-ROM drive, a memory card, etc. As another example, the software components may be loaded into the memory (710) via a communication module (730) other than a computer-readable recording medium. For example, at least one program may be loaded into the memory (710) based on a computer program that is installed by files provided by developers or a file distribution system that distributes installation files of applications via the communication module (730).

[0084] The processor (720) may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input / output operations. The instructions may be provided to another user terminal (not shown) or another external system via the memory (710) or the communication module (730).

[0085] The communication module (730) may provide a configuration or function for a user terminal (not shown) and a computing device (200) to communicate with each other via a network, and may provide a configuration or function for the computing device (200) to communicate with an external system (e.g., a separate cloud system, etc.). For example, control signals, commands, data, etc. provided under the control of the processor (720) of the computing device (200) may be transmitted to the user terminal and / or the external system via the communication module (730) and the network via the communication module of the user terminal and / or the external system.

[0086] In addition, the input / output interface (740) of the computing device (200) may be a means for interfacing with a device (not shown) for input or output that is connected to the computing device (200) or that the computing device (200) may include. In FIG. 7, the input / output interface (740) is illustrated as an element configured separately from the processor (720), but is not limited thereto, and the input / output interface (740) may be configured to be included in the processor (720). The computing device (200) may include more components than those illustrated in FIG. 7. However, there is no need to explicitly illustrate most of the conventional components.

[0087] The processor (720) of the computing device (200) may be configured to manage, process and / or store information and / or data received from multiple user terminals and / or multiple external systems.

[0088]

[0089] Figure 8 is a drawing for explaining the effect according to the present disclosure.

[0090] Referring to Figure 8, the technique according to one embodiment of the present disclosure, represented by a graph with GAN+SMOTE, demonstrates higher accuracy compared to conventional techniques that separately apply GAN and SMOTE. Therefore, it can be confirmed that the recognition rate can be improved by utilizing the augmented data according to the present disclosure.

[0091]

[0092] The above-described methods and / or various embodiments may be realized by digital electronic circuits, computer hardware, firmware, software, and / or a combination thereof. Various embodiments of the present disclosure may be implemented as a computer program that is executed by a data processing device, for example, one or more programmable processors and / or one or more computing devices, or stored on a computer-readable recording medium and / or a computer-readable recording medium. The above-described computer program may be written in any form of programming language, including a compiled language or an interpreted language, and may be distributed in any form, such as a standalone program, a module, a subroutine, etc. The computer program may be distributed through a single computing device, multiple computing devices connected through the same network, and / or multiple computing devices distributed to be connected through multiple different networks.

[0093] The methods and / or various embodiments described above may be performed by one or more processors configured to execute one or more computer programs that process, store, and / or manage any function, function, etc. by operating on the basis of input data or generating output data. For example, the methods and / or various embodiments of the present disclosure may be performed by special purpose logic circuits such as Field Programmable Gate Arrays (FPGAs) or Application Specific Integrated Circuits (ASICs), and an apparatus and / or system for performing the methods and / or embodiments of the present disclosure may be implemented as special purpose logic circuits such as FPGAs or ASICs.

[0094] The one or more processors executing the computer program may include a general-purpose or special-purpose microprocessor and / or one or more processors of any type of digital computing device. The processor may receive instructions and / or data from each of read-only memory and random-access memory, or may receive instructions and / or data from the read-only memory and the random-access memory. In the present invention, components of a computing device performing the methods and / or embodiments may include one or more processors for executing instructions, and one or more memory devices for storing instructions and / or data.

[0095] According to one embodiment, the computing device can transmit and receive data to and from one or more mass storage devices for storing data. For example, the computing device can receive and / or transfer data from a magnetic disc or an optical disc. A computer-readable storage medium suitable for storing instructions and / or data associated with a computer program may include, but is not limited to, any form of non-volatile memory, including semiconductor memory devices such as Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable PROM (EEPROM), flash memory devices, and the like. For example, the computer-readable storage medium may include a magnetic disk such as an internal hard disk or a removable disk, a magneto-optical disk, a CD-ROM, and a DVD-ROM disk.

[0096] To provide interaction with a user, a computing device may include, but is not limited to, a display device (e.g., a cathode ray tube (CRT), a liquid crystal display (LCD), etc.) for providing or displaying information to the user, and a pointing device (e.g., a keyboard, a mouse, a trackball, etc.) for allowing the user to provide input and / or commands to the computing device. That is, the computing device may further include any other types of devices for providing interaction with the user. For example, the computing device may provide any form of sensory feedback to the user, including visual feedback, auditory feedback, and / or tactile feedback, for interaction with the user. In this regard, the user may provide input to the computing device through various gestures, such as visual, vocal, or motion.

[0097] In the present invention, various embodiments may be implemented in a computing system that includes backend components (e.g., a data server), middleware components (e.g., an application server), and / or front-end components. In this case, the components may be interconnected via any form or medium of digital data communication, such as a communications network. For example, the communications network may include a Local Area Network (LAN), a Wide Area Network (WAN), etc.

[0098] A computing device based on the exemplary embodiments described herein may be implemented using hardware and / or software configured to interact with a user, including a user device, a user interface (UI) device, a user terminal, or a client device. For example, the computing device may include a portable computing device, such as a laptop computer. Additionally or alternatively, the computing device may include, but is not limited to, Personal Digital Assistants (PDAs), tablet PCs, game consoles, wearable devices, Internet of Things (IoT) devices, virtual reality (VR) devices, augmented reality (AR) devices, and the like. The computing device may further include other types of devices configured to interact with a user. Furthermore, the computing device may include a portable communication device (e.g., a mobile phone, a smart phone, a wireless cellular phone, etc.) suitable for wireless communication over a network, such as a mobile communication network. The computing device may be configured to communicate wirelessly with a network server using wireless communication technologies and / or protocols, such as Radio Frequency (RF), Microwave Frequency (MWF), and / or Infrared Ray Frequency (IRF).

[0099] The various embodiments of the present invention, including specific structural and functional details, are exemplary. Therefore, the embodiments of the present disclosure are not limited to those described above and may be implemented in various other forms. Furthermore, the terminology used herein is intended to describe certain embodiments and is not intended to limit the embodiments. For example, singular terms and the above may be interpreted to include plural forms, unless the context clearly dictates otherwise.

[0100] In the present invention, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by those of ordinary skill in the art to which they pertain. Furthermore, commonly used terms, such as terms defined in dictionaries, should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology.

[0101] While the present invention has been described in connection with certain embodiments herein, various modifications and variations can be made without departing from the scope of the present disclosure, which would be apparent to those skilled in the art. Furthermore, such modifications and variations are intended to fall within the scope of the claims appended to this specification.

Claims

1. A method for data augmentation of an ultra-wideband (UWB) signal based on an artificial intelligence model performed by at least one processor, A step of receiving a first UWB signal and a second UWB signal; A step of generating a third UWB signal based on the first UWB signal and the second UWB signal using an interpolation algorithm; and A step of providing at least one of the first UWB signal, the second UWB signal and the third UWB signal to a learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal; A method for data augmentation of UWB signals based on an artificial intelligence model including:

2. In paragraph 1, The step of generating the third UWB signal is: A step of generating the third UWB signal based on third label data located between first labeled data corresponding to the first UWB signal and second labeled data corresponding to the second UWB signal in a data space; A method for data augmentation of UWB signals based on an artificial intelligence model including:

3. In paragraph 1, The step of generating the third UWB signal is: A step of generating the third UWB signal using a SMOTE (synthetic minority over-sampling technique) algorithm; A method for data augmentation of UWB signals based on an artificial intelligence model including:

4. In paragraph 1, The above artificial intelligence model is, A method for data augmentation of UWB signals based on an artificial intelligence model, which is a conditional generative adversarial net (CGAN)-based model including a generator and a discriminator.

5. In paragraph 4, The above generator is a data augmentation method of UWB signals based on an artificial intelligence model having a 1D U-Net structure.

6. In paragraph 4, The above constructor, In response to receiving a sampling vector and a condition vector corresponding to a class of UWB signals, the method is trained to generate a fake UWB signal having a distribution similar to at least one of the first UWB signal, the second UWB signal and the third UWB signal. The above discriminator is, A data augmentation method of a UWB signal based on an artificial intelligence model, which learns to determine whether the fake UWB signal is genuine or not in response to receiving the generated fake UWB signal.

7. In paragraph 6, A method for data augmentation of UWB signals based on an artificial intelligence model, wherein the discriminator comprises an auxiliary classifier for classifying the class of the fake UWB signal.

8. In paragraph 1, A data augmentation method of an artificial intelligence model-based UWB signal, wherein the first UWB signal and the second UWB signal are generated by measuring an object at the same angle.

9. A computer-readable, non-transitory recording medium having recorded thereon a program for executing the data augmentation method of a UWB signal based on an artificial intelligence model described in Article 1.

10. As a computing device, Communication module; memory; and At least one processor connected to said memory and configured to execute at least one computer-readable program contained in said memory Including, At least one of the above programs, Receives a first UWB signal and a second UWB signal, Generating a third UWB signal based on the first UWB signal and the second UWB signal using an interpolation algorithm, A computing device comprising instructions for providing at least one of the first UWB signal, the second UWB signal and the third UWB signal to a learned artificial intelligence model to generate a fourth UWB signal having a distribution associated with the at least one signal.

Images